framework,version,device,op_name,kernel_source,model,architecture,mla_dtype,kv_cache_dtype,gemm_type,num_heads,batch_size,isl,tp_size,step,latency
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,1,0.1525
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,4,0.1523
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,8,0.1522
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,16,0.1526
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,32,0.1531
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,64,0.1524
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,128,0.1563
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,256,0.1585
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,512,0.1573
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,1024,0.1581
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,2048,0.1663
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,4096,0.1682
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,8192,0.1686
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1,1,1,16384,0.1710
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,1,0.1578
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,4,0.1561
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,8,0.1562
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,16,0.1563
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,32,0.1567
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,64,0.1568
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,128,0.1588
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,256,0.1641
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,512,0.1616
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,1024,0.1649
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,2048,0.1743
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,4096,0.1766
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,8192,0.1772
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,2,1,1,16384,0.1792
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,1,0.1608
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,4,0.1605
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,8,0.1607
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,16,0.1604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,32,0.1607
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,64,0.1608
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,128,0.1635
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,256,0.1704
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,512,0.1695
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,1024,0.1684
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,2048,0.1790
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,4096,0.1866
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,8192,0.1877
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,4,1,1,16384,0.1917
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,1,0.1338
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,4,0.1342
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,8,0.1352
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,16,0.1345
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,32,0.1349
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,64,0.1352
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,128,0.1379
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,256,0.1419
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,512,0.1460
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,1024,0.1519
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,2048,0.1621
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,4096,0.1643
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,8192,0.1659
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,8,1,1,16384,0.1704
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,1,0.1428
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,4,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,8,0.1434
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,16,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,32,0.1438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,64,0.1432
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,128,0.1460
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,256,0.1471
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,512,0.1491
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,1024,0.1520
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,2048,0.1633
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,4096,0.1665
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,8192,0.1700
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,16,1,1,16384,0.1748
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,1,0.1458
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,4,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,8,0.1460
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,16,0.1459
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,32,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,64,0.1465
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,128,0.1493
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,256,0.1533
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,512,0.1551
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,1024,0.1598
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,2048,0.1766
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,32,1,1,4096,0.1807
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,1,0.1528
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,4,0.1532
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,8,0.1539
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,16,0.1528
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,32,0.1533
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,64,0.1534
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,128,0.1576
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,256,0.1594
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,512,0.1641
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,1024,0.1730
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,2048,0.1975
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,64,1,1,4096,0.2022
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,1,0.1717
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,4,0.1708
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,8,0.1714
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,16,0.1718
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,32,0.1719
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,64,0.1719
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,128,0.1779
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,256,0.1827
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,512,0.1918
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,1024,0.2107
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,128,1,1,2048,0.2510
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,1,0.2145
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,4,0.2146
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,8,0.2146
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,16,0.2145
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,32,0.2147
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,64,0.2151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,128,0.2254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,256,0.2318
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,512,0.2531
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,256,1,1,1024,0.2918
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,1,0.2867
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,4,0.2859
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,8,0.2871
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,16,0.2872
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,32,0.2869
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,64,0.2872
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,128,0.3064
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,256,0.3214
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,512,1,1,512,0.3553
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1024,1,1,1,0.4706
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1024,1,1,4,0.4696
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1024,1,1,8,0.4702
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1024,1,1,16,0.4703
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1024,1,1,32,0.4705
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1024,1,1,64,0.4724
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1024,1,1,128,0.5015
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,128,1024,1,1,256,0.5284
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,1,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,4,0.1262
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,8,0.1363
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,16,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,32,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,64,0.1360
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,128,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,256,0.1405
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,512,0.1416
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,1024,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,2048,0.1453
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,4096,0.1524
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,8192,0.1533
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1,1,1,16384,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,1,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,4,0.1383
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,8,0.1386
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,16,0.1385
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,32,0.1387
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,64,0.1399
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,128,0.1413
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,256,0.1429
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,512,0.1439
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,1024,0.1451
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,2048,0.1523
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,4096,0.1542
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,8192,0.1559
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,2,1,1,16384,0.1579
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,1,0.1395
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,4,0.1373
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,8,0.1393
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,16,0.1378
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,32,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,64,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,128,0.1413
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,256,0.1451
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,512,0.1446
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,1024,0.1447
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,2048,0.1549
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,4096,0.1576
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,8192,0.1579
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,4,1,1,16384,0.1622
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,1,0.1382
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,4,0.1390
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,8,0.1394
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,16,0.1378
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,32,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,64,0.1383
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,128,0.1414
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,256,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,512,0.1447
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,1024,0.1495
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,2048,0.1605
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,4096,0.1625
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,8192,0.1641
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,8,1,1,16384,0.1691
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,1,0.1285
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,4,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,8,0.1290
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,16,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,32,0.1283
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,64,0.1295
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,128,0.1320
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,256,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,512,0.1399
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,1024,0.1479
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,2048,0.1569
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,4096,0.1610
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,8192,0.1630
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,16,1,1,16384,0.1687
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,1,0.1311
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,4,0.1295
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,8,0.1309
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,16,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,32,0.1301
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,64,0.1310
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,128,0.1331
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,256,0.1424
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,512,0.1461
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,1024,0.1518
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,2048,0.1670
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,32,1,1,4096,0.1690
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,1,0.1368
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,4,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,8,0.1388
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,16,0.1374
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,32,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,64,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,128,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,256,0.1521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,512,0.1569
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,1024,0.1642
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,2048,0.1882
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,64,1,1,4096,0.1948
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,1,0.1473
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,4,0.1478
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,8,0.1475
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,16,0.1487
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,32,0.1478
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,64,0.1473
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,128,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,256,0.1573
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,512,0.1673
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,1024,0.1850
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,128,1,1,2048,0.2223
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,1,0.1779
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,4,0.1772
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,8,0.1771
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,16,0.1789
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,32,0.1770
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,64,0.1782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,128,0.1863
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,256,0.1966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,512,0.2144
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,256,1,1,1024,0.2544
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,1,0.2285
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,4,0.2276
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,8,0.2275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,16,0.2278
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,32,0.2271
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,64,0.2290
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,128,0.2464
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,256,0.2653
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,512,1,1,512,0.3020
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1024,1,1,1,0.3298
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1024,1,1,4,0.3307
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1024,1,1,8,0.3322
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1024,1,1,16,0.3319
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1024,1,1,32,0.3324
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1024,1,1,64,0.3345
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1024,1,1,128,0.3674
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,64,1024,1,1,256,0.4012
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1,1,1,1,0.0989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1,1,1,4,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1,1,1,8,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1,1,1,16,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1,1,1,64,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1,1,1,128,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,2,1,1,1,0.1083
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,2,1,1,4,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,2,1,1,8,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,2,1,1,16,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,2,1,1,32,0.1088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,2,1,1,64,0.1072
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,2,1,1,128,0.1108
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,4,1,1,1,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,4,1,1,4,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,4,1,1,8,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,4,1,1,16,0.1096
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,4,1,1,32,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,4,1,1,64,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,4,1,1,128,0.1127
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,8,1,1,1,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,8,1,1,4,0.1092
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,8,1,1,8,0.1109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,8,1,1,16,0.1103
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,8,1,1,32,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,8,1,1,64,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,8,1,1,128,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,16,1,1,1,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,16,1,1,4,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,16,1,1,8,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,16,1,1,16,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,16,1,1,32,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,16,1,1,64,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,16,1,1,128,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,32,1,1,1,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,32,1,1,4,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,32,1,1,8,0.1084
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,32,1,1,16,0.1092
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,32,1,1,32,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,32,1,1,64,0.1088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,32,1,1,128,0.1127
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,64,1,1,1,0.1120
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,64,1,1,4,0.1117
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,64,1,1,8,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,64,1,1,16,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,64,1,1,32,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,64,1,1,64,0.1184
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,64,1,1,128,0.1193
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,128,1,1,1,0.1289
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,128,1,1,4,0.1291
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,128,1,1,8,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,128,1,1,16,0.1289
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,128,1,1,32,0.1292
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,128,1,1,64,0.1315
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,128,1,1,128,0.1379
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,256,1,1,1,0.1514
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,256,1,1,4,0.1534
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,256,1,1,8,0.1534
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,256,1,1,16,0.1538
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,256,1,1,32,0.1536
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,256,1,1,64,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,256,1,1,128,0.1659
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,512,1,1,1,0.1954
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,512,1,1,4,0.1947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,512,1,1,8,0.1972
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,512,1,1,16,0.1950
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,512,1,1,32,0.1967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,512,1,1,64,0.1974
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,512,1,1,128,0.2145
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1024,1,1,1,0.2720
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1024,1,1,4,0.2719
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1024,1,1,8,0.2734
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1024,1,1,16,0.2717
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1024,1,1,32,0.2724
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1024,1,1,64,0.2740
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,32,1024,1,1,128,0.3061
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1,1,1,1,0.0905
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1,1,1,4,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1,1,1,8,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1,1,1,16,0.0949
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1,1,1,32,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1,1,1,64,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1,1,1,128,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,2,1,1,1,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,2,1,1,4,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,2,1,1,8,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,2,1,1,16,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,2,1,1,32,0.1025
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,2,1,1,64,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,2,1,1,128,0.1037
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,4,1,1,1,0.1036
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,4,1,1,4,0.1045
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,4,1,1,8,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,4,1,1,16,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,4,1,1,32,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,4,1,1,64,0.1032
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,4,1,1,128,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,8,1,1,1,0.1045
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,8,1,1,4,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,8,1,1,8,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,8,1,1,16,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,8,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,8,1,1,64,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,8,1,1,128,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,16,1,1,1,0.1055
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,16,1,1,4,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,16,1,1,8,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,16,1,1,16,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,16,1,1,32,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,16,1,1,64,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,16,1,1,128,0.1031
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,32,1,1,1,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,32,1,1,4,0.1074
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,32,1,1,8,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,32,1,1,16,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,32,1,1,32,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,32,1,1,64,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,32,1,1,128,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,64,1,1,1,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,64,1,1,4,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,64,1,1,8,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,64,1,1,16,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,64,1,1,32,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,64,1,1,64,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,64,1,1,128,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,128,1,1,1,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,128,1,1,4,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,128,1,1,8,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,128,1,1,16,0.1103
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,128,1,1,32,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,128,1,1,64,0.1108
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,128,1,1,128,0.1191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,256,1,1,1,0.1304
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,256,1,1,4,0.1332
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,256,1,1,8,0.1315
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,256,1,1,16,0.1323
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,256,1,1,32,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,256,1,1,64,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,256,1,1,128,0.1503
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,512,1,1,1,0.1764
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,512,1,1,4,0.1747
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,512,1,1,8,0.1766
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,512,1,1,16,0.1756
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,512,1,1,32,0.1768
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,512,1,1,64,0.1786
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,512,1,1,128,0.1955
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1024,1,1,1,0.2431
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1024,1,1,4,0.2438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1024,1,1,8,0.2441
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1024,1,1,16,0.2438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1024,1,1,32,0.2433
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1024,1,1,64,0.2461
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,16,1024,1,1,128,0.2761
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1,1,1,1,0.0865
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1,1,1,4,0.0864
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1,1,1,8,0.0925
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1,1,1,16,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1,1,1,32,0.0905
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1,1,1,64,0.0925
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1,1,1,128,0.0945
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,2,1,1,1,0.0958
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,2,1,1,4,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,2,1,1,8,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,2,1,1,16,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,2,1,1,32,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,2,1,1,64,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,2,1,1,128,0.0977
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,4,1,1,1,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,4,1,1,4,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,4,1,1,8,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,4,1,1,16,0.0982
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,4,1,1,32,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,4,1,1,64,0.0990
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,4,1,1,128,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,8,1,1,1,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,8,1,1,4,0.0972
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,8,1,1,8,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,8,1,1,16,0.0984
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,8,1,1,32,0.0975
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,8,1,1,64,0.0974
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,8,1,1,128,0.0996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,16,1,1,1,0.1012
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,16,1,1,4,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,16,1,1,8,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,16,1,1,16,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,16,1,1,32,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,16,1,1,64,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,16,1,1,128,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,32,1,1,1,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,32,1,1,4,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,32,1,1,8,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,32,1,1,16,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,32,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,32,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,32,1,1,128,0.1031
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,64,1,1,1,0.1031
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,64,1,1,4,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,64,1,1,8,0.1036
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,64,1,1,16,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,64,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,64,1,1,64,0.1032
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,64,1,1,128,0.1052
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,128,1,1,1,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,128,1,1,4,0.1055
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,128,1,1,8,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,128,1,1,16,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,128,1,1,32,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,128,1,1,64,0.1055
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,128,1,1,128,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,256,1,1,1,0.1213
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,256,1,1,4,0.1216
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,256,1,1,8,0.1220
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,256,1,1,16,0.1227
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,256,1,1,32,0.1214
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,256,1,1,64,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,256,1,1,128,0.1404
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,512,1,1,1,0.1627
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,512,1,1,4,0.1619
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,512,1,1,8,0.1650
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,512,1,1,16,0.1664
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,512,1,1,32,0.1686
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,512,1,1,64,0.1705
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,512,1,1,128,0.1885
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1024,1,1,1,0.2272
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1024,1,1,4,0.2278
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1024,1,1,8,0.2287
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1024,1,1,16,0.2276
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1024,1,1,32,0.2278
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1024,1,1,64,0.2313
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,float16,8,1024,1,1,128,0.2604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,1,0.1513
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,4,0.1492
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,8,0.1599
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,16,0.1604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,32,0.1608
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,64,0.1607
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,128,0.1604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,256,0.1647
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,512,0.1626
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,1024,0.1639
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,2048,0.1705
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,4096,0.1726
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,8192,0.1758
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1,1,1,16384,0.1758
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,1,0.1647
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,4,0.1651
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,8,0.1652
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,16,0.1652
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,32,0.1657
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,64,0.1659
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,128,0.1652
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,256,0.1690
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,512,0.1691
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,1024,0.1703
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,2048,0.1788
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,4096,0.1817
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,8192,0.1830
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,2,1,1,16384,0.1843
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,1,0.1703
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,4,0.1707
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,8,0.1704
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,16,0.1701
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,32,0.1701
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,64,0.1706
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,128,0.1710
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,256,0.1717
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,512,0.1736
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,1024,0.1709
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,2048,0.1809
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,4096,0.1834
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,8192,0.1852
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,4,1,1,16384,0.1880
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,1,0.1428
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,4,0.1433
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,8,0.1438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,16,0.1432
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,32,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,64,0.1426
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,128,0.1434
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,256,0.1483
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,512,0.1488
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,1024,0.1543
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,2048,0.1631
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,4096,0.1668
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,8192,0.1675
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,8,1,1,16384,0.1721
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,1,0.1480
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,4,0.1470
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,8,0.1478
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,16,0.1496
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,32,0.1482
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,64,0.1479
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,128,0.1498
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,256,0.1536
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,512,0.1543
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,1024,0.1563
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,2048,0.1651
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,4096,0.1703
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,8192,0.1713
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,16,1,1,16384,0.1773
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,1,0.1495
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,4,0.1503
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,8,0.1508
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,16,0.1499
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,32,0.1498
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,64,0.1516
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,128,0.1518
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,256,0.1587
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,512,0.1602
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,1024,0.1618
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,2048,0.1744
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,32,1,1,4096,0.1780
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,1,0.1570
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,4,0.1566
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,8,0.1573
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,16,0.1586
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,32,0.1566
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,64,0.1568
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,128,0.1588
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,256,0.1598
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,512,0.1625
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,1024,0.1670
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,2048,0.1844
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,64,1,1,4096,0.1875
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,1,0.1727
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,4,0.1733
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,8,0.1740
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,16,0.1729
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,32,0.1733
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,64,0.1737
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,128,0.1764
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,256,0.1782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,512,0.1815
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,1024,0.1911
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,128,1,1,2048,0.2156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,1,0.2155
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,4,0.2155
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,8,0.2146
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,16,0.2153
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,32,0.2162
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,64,0.2157
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,128,0.2179
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,256,0.2239
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,512,0.2327
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,256,1,1,1024,0.2525
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,1,0.2792
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,4,0.2792
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,8,0.2806
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,16,0.2794
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,32,0.2789
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,64,0.2796
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,128,0.2842
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,256,0.2956
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,512,1,1,512,0.3115
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1024,1,1,1,0.4483
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1024,1,1,4,0.4482
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1024,1,1,8,0.4482
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1024,1,1,16,0.4476
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1024,1,1,32,0.4508
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1024,1,1,64,0.4501
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1024,1,1,128,0.4605
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,128,1024,1,1,256,0.4791
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,1,0.1232
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,4,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,8,0.1331
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,16,0.1338
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,32,0.1338
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,64,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,128,0.1338
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,256,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,512,0.1363
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,1024,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,2048,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,4096,0.1463
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,8192,0.1498
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1,1,1,16384,0.1501
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,1,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,4,0.1366
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,8,0.1383
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,16,0.1373
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,32,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,64,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,128,0.1372
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,256,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,512,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,1024,0.1393
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,2048,0.1473
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,4096,0.1505
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,8192,0.1517
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,2,1,1,16384,0.1523
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,1,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,4,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,8,0.1369
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,16,0.1370
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,32,0.1362
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,64,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,128,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,256,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,512,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,1024,0.1398
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,2048,0.1495
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,4096,0.1516
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,8192,0.1546
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,4,1,1,16384,0.1569
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,1,0.1346
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,4,0.1358
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,8,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,16,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,32,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,64,0.1360
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,128,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,256,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,512,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,1024,0.1417
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,2048,0.1518
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,4096,0.1537
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,8192,0.1551
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,8,1,1,16384,0.1598
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,1,0.1261
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,4,0.1250
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,8,0.1259
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,16,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,32,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,64,0.1242
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,128,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,256,0.1313
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,512,0.1319
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,1024,0.1386
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,2048,0.1480
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,4096,0.1518
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,8192,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,16,1,1,16384,0.1600
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,1,0.1270
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,4,0.1271
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,8,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,16,0.1275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,32,0.1271
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,64,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,128,0.1285
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,256,0.1342
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,512,0.1378
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,1024,0.1407
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,2048,0.1528
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,32,1,1,4096,0.1562
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,1,0.1339
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,4,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,8,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,16,0.1339
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,32,0.1337
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,64,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,128,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,256,0.1438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,512,0.1461
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,1024,0.1506
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,2048,0.1679
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,64,1,1,4096,0.1715
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,1,0.1408
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,4,0.1413
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,8,0.1412
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,16,0.1422
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,32,0.1414
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,64,0.1422
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,128,0.1461
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,256,0.1476
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,512,0.1520
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,1024,0.1615
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,128,1,1,2048,0.1854
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,1,0.1716
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,4,0.1709
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,8,0.1696
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,16,0.1719
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,32,0.1713
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,64,0.1704
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,128,0.1754
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,256,0.1797
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,512,0.1892
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,256,1,1,1024,0.2088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,1,0.2173
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,4,0.2183
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,8,0.2194
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,16,0.2187
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,32,0.2183
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,64,0.2191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,128,0.2285
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,256,0.2389
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,512,1,1,512,0.2566
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1024,1,1,1,0.3169
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1024,1,1,4,0.3162
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1024,1,1,8,0.3165
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1024,1,1,16,0.3165
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1024,1,1,32,0.3167
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1024,1,1,64,0.3195
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1024,1,1,128,0.3337
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,64,1024,1,1,256,0.3494
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1,1,1,1,0.0969
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1,1,1,4,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1,1,1,8,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1,1,1,16,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1,1,1,32,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1,1,1,64,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1,1,1,128,0.1046
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,2,1,1,1,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,2,1,1,4,0.1071
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,2,1,1,8,0.1071
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,2,1,1,16,0.1063
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,2,1,1,32,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,2,1,1,64,0.1073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,2,1,1,128,0.1073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,4,1,1,1,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,4,1,1,4,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,4,1,1,8,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,4,1,1,16,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,4,1,1,32,0.1101
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,4,1,1,64,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,4,1,1,128,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,8,1,1,1,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,8,1,1,4,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,8,1,1,8,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,8,1,1,16,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,8,1,1,32,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,8,1,1,64,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,8,1,1,128,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,16,1,1,1,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,16,1,1,4,0.1097
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,16,1,1,8,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,16,1,1,16,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,16,1,1,32,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,16,1,1,64,0.1092
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,16,1,1,128,0.1095
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,32,1,1,1,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,32,1,1,4,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,32,1,1,8,0.1056
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,32,1,1,16,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,32,1,1,32,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,32,1,1,64,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,32,1,1,128,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,64,1,1,1,0.1074
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,64,1,1,4,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,64,1,1,8,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,64,1,1,16,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,64,1,1,32,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,64,1,1,64,0.1101
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,64,1,1,128,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,128,1,1,1,0.1157
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,128,1,1,4,0.1162
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,128,1,1,8,0.1154
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,128,1,1,16,0.1144
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,128,1,1,32,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,128,1,1,64,0.1173
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,128,1,1,128,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,256,1,1,1,0.1435
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,256,1,1,4,0.1441
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,256,1,1,8,0.1439
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,256,1,1,16,0.1424
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,256,1,1,32,0.1447
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,256,1,1,64,0.1442
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,256,1,1,128,0.1521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,512,1,1,1,0.1844
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,512,1,1,4,0.1846
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,512,1,1,8,0.1840
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,512,1,1,16,0.1834
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,512,1,1,32,0.1856
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,512,1,1,64,0.1848
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,512,1,1,128,0.1966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1024,1,1,1,0.2568
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1024,1,1,4,0.2571
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1024,1,1,8,0.2568
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1024,1,1,16,0.2570
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1024,1,1,32,0.2587
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1024,1,1,64,0.2581
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,32,1024,1,1,128,0.2762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1,1,1,1,0.0891
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1,1,1,4,0.0888
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1,1,1,8,0.0950
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1,1,1,16,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1,1,1,32,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1,1,1,64,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1,1,1,128,0.0965
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,2,1,1,1,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,2,1,1,4,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,2,1,1,8,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,2,1,1,16,0.0993
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,2,1,1,32,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,2,1,1,64,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,2,1,1,128,0.1015
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,4,1,1,1,0.1024
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,4,1,1,4,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,4,1,1,8,0.1017
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,4,1,1,16,0.1014
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,4,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,4,1,1,64,0.1021
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,4,1,1,128,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,8,1,1,1,0.1012
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,8,1,1,4,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,8,1,1,8,0.1010
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,8,1,1,16,0.1015
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,8,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,8,1,1,64,0.1019
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,8,1,1,128,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,16,1,1,1,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,16,1,1,4,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,16,1,1,8,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,16,1,1,16,0.1017
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,16,1,1,32,0.1026
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,16,1,1,64,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,16,1,1,128,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,32,1,1,1,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,32,1,1,4,0.1033
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,32,1,1,8,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,32,1,1,16,0.1035
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,32,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,32,1,1,64,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,32,1,1,128,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,64,1,1,1,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,64,1,1,4,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,64,1,1,8,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,64,1,1,16,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,64,1,1,32,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,64,1,1,64,0.1031
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,64,1,1,128,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,128,1,1,1,0.1037
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,128,1,1,4,0.1052
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,128,1,1,8,0.1032
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,128,1,1,16,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,128,1,1,32,0.1060
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,128,1,1,64,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,128,1,1,128,0.1076
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,256,1,1,1,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,256,1,1,4,0.1295
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,256,1,1,8,0.1238
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,256,1,1,16,0.1277
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,256,1,1,32,0.1272
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,256,1,1,64,0.1290
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,256,1,1,128,0.1358
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,512,1,1,1,0.1645
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,512,1,1,4,0.1645
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,512,1,1,8,0.1641
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,512,1,1,16,0.1636
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,512,1,1,32,0.1664
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,512,1,1,64,0.1648
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,512,1,1,128,0.1766
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1024,1,1,1,0.2270
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1024,1,1,4,0.2269
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1024,1,1,8,0.2264
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1024,1,1,16,0.2267
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1024,1,1,32,0.2275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1024,1,1,64,0.2275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,16,1024,1,1,128,0.2463
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1,1,1,1,0.0848
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1,1,1,4,0.0845
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1,1,1,8,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1,1,1,16,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1,1,1,32,0.0905
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1,1,1,64,0.0909
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1,1,1,128,0.0925
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,2,1,1,1,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,2,1,1,4,0.0955
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,2,1,1,8,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,2,1,1,16,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,2,1,1,32,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,2,1,1,64,0.0959
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,2,1,1,128,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,4,1,1,1,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,4,1,1,4,0.0976
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,4,1,1,8,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,4,1,1,16,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,4,1,1,32,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,4,1,1,64,0.0976
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,4,1,1,128,0.0974
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,8,1,1,1,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,8,1,1,4,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,8,1,1,8,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,8,1,1,16,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,8,1,1,32,0.0971
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,8,1,1,64,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,8,1,1,128,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,16,1,1,1,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,16,1,1,4,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,16,1,1,8,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,16,1,1,16,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,16,1,1,32,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,16,1,1,64,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,16,1,1,128,0.1000
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,32,1,1,1,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,32,1,1,4,0.0991
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,32,1,1,8,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,32,1,1,16,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,32,1,1,32,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,32,1,1,64,0.0994
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,32,1,1,128,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,64,1,1,1,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,64,1,1,4,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,64,1,1,8,0.1017
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,64,1,1,16,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,64,1,1,32,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,64,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,64,1,1,128,0.1015
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,128,1,1,1,0.1025
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,128,1,1,4,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,128,1,1,8,0.1022
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,128,1,1,16,0.1011
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,128,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,128,1,1,64,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,128,1,1,128,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,256,1,1,1,0.1152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,256,1,1,4,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,256,1,1,8,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,256,1,1,16,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,256,1,1,32,0.1160
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,256,1,1,64,0.1192
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,256,1,1,128,0.1224
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,512,1,1,1,0.1512
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,512,1,1,4,0.1519
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,512,1,1,8,0.1510
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,512,1,1,16,0.1518
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,512,1,1,32,0.1536
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,512,1,1,64,0.1563
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,512,1,1,128,0.1688
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1024,1,1,1,0.2081
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1024,1,1,4,0.2087
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1024,1,1,8,0.2085
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1024,1,1,16,0.2090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1024,1,1,32,0.2100
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1024,1,1,64,0.2097
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,float16,8,1024,1,1,128,0.2294
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,1,0.1525
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,4,0.1521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,8,0.1625
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,16,0.1629
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,32,0.1619
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,64,0.1621
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,128,0.1662
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,256,0.1680
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,512,0.1684
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,1024,0.1679
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,2048,0.1759
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,4096,0.1783
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,8192,0.1803
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1,1,1,16384,0.1805
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,1,0.1682
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,4,0.1674
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,8,0.1673
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,16,0.1674
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,32,0.1685
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,64,0.1688
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,128,0.1688
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,256,0.1753
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,512,0.1723
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,1024,0.1763
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,2048,0.1838
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,4096,0.1862
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,8192,0.1865
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,2,1,1,16384,0.1895
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,1,0.1728
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,4,0.1704
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,8,0.1723
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,16,0.1711
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,32,0.1723
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,64,0.1721
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,128,0.1736
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,256,0.1828
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,512,0.1793
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,1024,0.1797
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,2048,0.1893
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,4096,0.1977
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,8192,0.1985
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,4,1,1,16384,0.2021
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,1,0.1452
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,4,0.1440
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,8,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,16,0.1451
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,32,0.1460
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,64,0.1459
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,128,0.1478
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,256,0.1535
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,512,0.1565
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,1024,0.1638
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,2048,0.1725
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,4096,0.1752
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,8192,0.1766
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,8,1,1,16384,0.1820
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,1,0.1539
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,4,0.1537
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,8,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,16,0.1540
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,32,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,64,0.1552
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,128,0.1566
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,256,0.1588
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,512,0.1597
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,1024,0.1645
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,2048,0.1738
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,4096,0.1786
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,8192,0.1811
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,16,1,1,16384,0.1867
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,1,0.1565
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,4,0.1560
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,8,0.1564
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,16,0.1561
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,32,0.1566
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,64,0.1579
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,128,0.1599
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,256,0.1645
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,512,0.1672
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,1024,0.1707
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,2048,0.1879
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,32,1,1,4096,0.1914
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,1,0.1639
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,4,0.1641
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,8,0.1656
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,16,0.1638
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,32,0.1642
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,64,0.1641
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,128,0.1690
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,256,0.1705
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,512,0.1746
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,1024,0.1849
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,2048,0.2080
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,64,1,1,4096,0.2135
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,1,0.1816
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,4,0.1832
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,8,0.1818
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,16,0.1826
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,32,0.1825
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,64,0.1825
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,128,0.1900
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,256,0.1935
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,512,0.2028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,1024,0.2221
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,128,1,1,2048,0.2620
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,1,0.2257
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,4,0.2259
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,8,0.2255
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,16,0.2258
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,32,0.2251
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,64,0.2262
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,128,0.2365
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,256,0.2424
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,512,0.2630
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,256,1,1,1024,0.3027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,1,0.2989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,4,0.2963
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,8,0.2976
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,16,0.2987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,32,0.2977
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,64,0.2981
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,128,0.3176
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,256,0.3314
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,512,1,1,512,0.3676
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1024,1,1,1,0.4797
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1024,1,1,4,0.4817
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1024,1,1,8,0.4815
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1024,1,1,16,0.4813
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1024,1,1,32,0.4819
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1024,1,1,64,0.4821
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1024,1,1,128,0.5115
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,128,1024,1,1,256,0.5390
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,1,0.1253
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,4,0.1262
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,8,0.1365
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,16,0.1374
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,32,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,64,0.1364
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,128,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,256,0.1409
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,512,0.1414
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,1024,0.1422
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,2048,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,4096,0.1524
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,8192,0.1535
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1,1,1,16384,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,1,0.1398
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,4,0.1390
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,8,0.1381
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,16,0.1393
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,32,0.1385
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,64,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,128,0.1416
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,256,0.1430
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,512,0.1440
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,1024,0.1452
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,2048,0.1520
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,4096,0.1545
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,8192,0.1559
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,2,1,1,16384,0.1579
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,1,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,4,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,8,0.1404
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,16,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,32,0.1398
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,64,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,128,0.1405
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,256,0.1453
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,512,0.1445
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,1024,0.1440
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,2048,0.1543
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,4096,0.1578
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,8192,0.1582
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,4,1,1,16384,0.1621
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,1,0.1389
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,4,0.1379
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,8,0.1385
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,16,0.1379
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,32,0.1399
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,64,0.1394
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,128,0.1408
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,256,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,512,0.1441
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,1024,0.1498
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,2048,0.1604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,4096,0.1631
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,8192,0.1643
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,8,1,1,16384,0.1688
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,1,0.1279
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,4,0.1278
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,8,0.1287
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,16,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,32,0.1283
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,64,0.1297
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,128,0.1312
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,256,0.1354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,512,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,1024,0.1475
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,2048,0.1569
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,4096,0.1605
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,8192,0.1633
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,16,1,1,16384,0.1691
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,1,0.1314
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,4,0.1297
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,8,0.1311
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,16,0.1291
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,32,0.1298
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,64,0.1309
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,128,0.1332
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,256,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,512,0.1459
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,1024,0.1511
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,2048,0.1668
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,32,1,1,4096,0.1688
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,1,0.1358
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,4,0.1368
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,8,0.1380
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,16,0.1366
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,32,0.1365
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,64,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,128,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,256,0.1516
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,512,0.1559
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,1024,0.1643
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,2048,0.1887
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,64,1,1,4096,0.1940
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,1,0.1476
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,4,0.1476
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,8,0.1477
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,16,0.1479
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,32,0.1473
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,64,0.1477
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,128,0.1537
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,256,0.1576
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,512,0.1678
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,1024,0.1850
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,128,1,1,2048,0.2220
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,1,0.1784
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,4,0.1776
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,8,0.1772
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,16,0.1784
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,32,0.1765
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,64,0.1785
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,128,0.1864
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,256,0.1962
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,512,0.2162
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,256,1,1,1024,0.2544
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,1,0.2286
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,4,0.2271
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,8,0.2276
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,16,0.2279
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,32,0.2271
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,64,0.2290
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,128,0.2461
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,256,0.2654
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,512,1,1,512,0.3014
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1024,1,1,1,0.3296
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1024,1,1,4,0.3314
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1024,1,1,8,0.3311
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1024,1,1,16,0.3324
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1024,1,1,32,0.3326
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1024,1,1,64,0.3350
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1024,1,1,128,0.3675
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,64,1024,1,1,256,0.4019
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1,1,1,1,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1,1,1,4,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1,1,1,8,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1,1,1,16,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1,1,1,64,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1,1,1,128,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,2,1,1,1,0.1081
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,2,1,1,4,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,2,1,1,8,0.1095
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,2,1,1,16,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,2,1,1,32,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,2,1,1,64,0.1080
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,2,1,1,128,0.1105
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,4,1,1,1,0.1109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,4,1,1,4,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,4,1,1,8,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,4,1,1,16,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,4,1,1,32,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,4,1,1,64,0.1109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,4,1,1,128,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,8,1,1,1,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,8,1,1,4,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,8,1,1,8,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,8,1,1,16,0.1100
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,8,1,1,32,0.1093
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,8,1,1,64,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,8,1,1,128,0.1114
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,16,1,1,1,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,16,1,1,4,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,16,1,1,8,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,16,1,1,16,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,16,1,1,32,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,16,1,1,64,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,16,1,1,128,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,32,1,1,1,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,32,1,1,4,0.1109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,32,1,1,8,0.1084
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,32,1,1,16,0.1092
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,32,1,1,32,0.1118
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,32,1,1,64,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,32,1,1,128,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,64,1,1,1,0.1115
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,64,1,1,4,0.1115
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,64,1,1,8,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,64,1,1,16,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,64,1,1,32,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,64,1,1,64,0.1180
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,64,1,1,128,0.1193
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,128,1,1,1,0.1292
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,128,1,1,4,0.1297
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,128,1,1,8,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,128,1,1,16,0.1288
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,128,1,1,32,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,128,1,1,64,0.1319
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,128,1,1,128,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,256,1,1,1,0.1517
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,256,1,1,4,0.1540
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,256,1,1,8,0.1539
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,256,1,1,16,0.1540
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,256,1,1,32,0.1530
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,256,1,1,64,0.1539
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,256,1,1,128,0.1654
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,512,1,1,1,0.1953
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,512,1,1,4,0.1951
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,512,1,1,8,0.1971
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,512,1,1,16,0.1950
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,512,1,1,32,0.1965
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,512,1,1,64,0.1978
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,512,1,1,128,0.2144
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1024,1,1,1,0.2718
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1024,1,1,4,0.2730
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1024,1,1,8,0.2731
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1024,1,1,16,0.2726
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1024,1,1,32,0.2722
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1024,1,1,64,0.2750
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,32,1024,1,1,128,0.3052
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1,1,1,1,0.0905
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1,1,1,4,0.0905
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1,1,1,8,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1,1,1,16,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1,1,1,32,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1,1,1,64,0.0965
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1,1,1,128,0.0985
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,2,1,1,1,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,2,1,1,4,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,2,1,1,8,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,2,1,1,16,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,2,1,1,32,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,2,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,2,1,1,128,0.1037
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,4,1,1,1,0.1043
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,4,1,1,4,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,4,1,1,8,0.1030
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,4,1,1,16,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,4,1,1,32,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,4,1,1,64,0.1037
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,4,1,1,128,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,8,1,1,1,0.1045
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,8,1,1,4,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,8,1,1,8,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,8,1,1,16,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,8,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,8,1,1,64,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,8,1,1,128,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,16,1,1,1,0.1059
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,16,1,1,4,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,16,1,1,8,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,16,1,1,16,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,16,1,1,32,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,16,1,1,64,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,16,1,1,128,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,32,1,1,1,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,32,1,1,4,0.1077
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,32,1,1,8,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,32,1,1,16,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,32,1,1,32,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,32,1,1,64,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,32,1,1,128,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,64,1,1,1,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,64,1,1,4,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,64,1,1,8,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,64,1,1,16,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,64,1,1,32,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,64,1,1,64,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,64,1,1,128,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,128,1,1,1,0.1113
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,128,1,1,4,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,128,1,1,8,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,128,1,1,16,0.1108
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,128,1,1,32,0.1112
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,128,1,1,64,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,128,1,1,128,0.1192
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,256,1,1,1,0.1302
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,256,1,1,4,0.1335
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,256,1,1,8,0.1319
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,256,1,1,16,0.1321
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,256,1,1,32,0.1340
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,256,1,1,64,0.1354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,256,1,1,128,0.1504
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,512,1,1,1,0.1764
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,512,1,1,4,0.1745
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,512,1,1,8,0.1768
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,512,1,1,16,0.1757
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,512,1,1,32,0.1771
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,512,1,1,64,0.1781
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,512,1,1,128,0.1954
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1024,1,1,1,0.2424
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1024,1,1,4,0.2438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1024,1,1,8,0.2443
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1024,1,1,16,0.2435
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1024,1,1,32,0.2433
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1024,1,1,64,0.2461
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,16,1024,1,1,128,0.2762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1,1,1,1,0.0864
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1,1,1,4,0.0865
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1,1,1,8,0.0926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1,1,1,16,0.0907
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1,1,1,32,0.0904
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1,1,1,64,0.0926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1,1,1,128,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,2,1,1,1,0.0961
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,2,1,1,4,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,2,1,1,8,0.0965
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,2,1,1,16,0.0965
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,2,1,1,32,0.0965
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,2,1,1,64,0.0949
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,2,1,1,128,0.0980
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,4,1,1,1,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,4,1,1,4,0.0990
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,4,1,1,8,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,4,1,1,16,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,4,1,1,32,0.0990
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,4,1,1,64,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,4,1,1,128,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,8,1,1,1,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,8,1,1,4,0.0975
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,8,1,1,8,0.0989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,8,1,1,16,0.0974
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,8,1,1,32,0.0976
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,8,1,1,64,0.0972
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,8,1,1,128,0.1001
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,16,1,1,1,0.1012
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,16,1,1,4,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,16,1,1,8,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,16,1,1,16,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,16,1,1,32,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,16,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,16,1,1,128,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,32,1,1,1,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,32,1,1,4,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,32,1,1,8,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,32,1,1,16,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,32,1,1,32,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,32,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,32,1,1,128,0.1030
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,64,1,1,1,0.1030
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,64,1,1,4,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,64,1,1,8,0.1041
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,64,1,1,16,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,64,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,64,1,1,64,0.1031
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,64,1,1,128,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,128,1,1,1,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,128,1,1,4,0.1055
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,128,1,1,8,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,128,1,1,16,0.1053
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,128,1,1,32,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,128,1,1,64,0.1058
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,128,1,1,128,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,256,1,1,1,0.1211
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,256,1,1,4,0.1216
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,256,1,1,8,0.1215
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,256,1,1,16,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,256,1,1,32,0.1213
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,256,1,1,64,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,256,1,1,128,0.1400
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,512,1,1,1,0.1627
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,512,1,1,4,0.1612
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,512,1,1,8,0.1655
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,512,1,1,16,0.1662
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,512,1,1,32,0.1690
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,512,1,1,64,0.1702
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,512,1,1,128,0.1881
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1024,1,1,1,0.2267
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1024,1,1,4,0.2280
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1024,1,1,8,0.2287
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1024,1,1,16,0.2277
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1024,1,1,32,0.2275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1024,1,1,64,0.2302
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,float16,fp8_block,8,1024,1,1,128,0.2598
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,1,0.1505
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,4,0.1497
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,8,0.1590
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,16,0.1602
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,32,0.1604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,64,0.1600
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,128,0.1605
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,256,0.1643
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,512,0.1625
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,1024,0.1635
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,2048,0.1709
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,4096,0.1730
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,8192,0.1751
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1,1,1,16384,0.1762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,1,0.1646
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,4,0.1648
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,8,0.1659
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,16,0.1647
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,32,0.1658
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,64,0.1658
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,128,0.1649
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,256,0.1693
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,512,0.1691
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,1024,0.1709
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,2048,0.1797
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,4096,0.1820
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,8192,0.1827
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,2,1,1,16384,0.1840
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,1,0.1707
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,4,0.1704
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,8,0.1707
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,16,0.1700
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,32,0.1702
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,64,0.1707
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,128,0.1715
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,256,0.1713
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,512,0.1735
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,1024,0.1711
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,2048,0.1808
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,4096,0.1833
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,8192,0.1849
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,4,1,1,16384,0.1879
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,1,0.1428
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,4,0.1433
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,8,0.1430
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,16,0.1426
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,32,0.1437
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,64,0.1430
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,128,0.1430
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,256,0.1480
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,512,0.1488
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,1024,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,2048,0.1638
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,4096,0.1664
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,8192,0.1675
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,8,1,1,16384,0.1720
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,1,0.1479
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,4,0.1469
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,8,0.1477
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,16,0.1488
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,32,0.1479
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,64,0.1478
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,128,0.1492
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,256,0.1534
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,512,0.1549
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,1024,0.1565
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,2048,0.1653
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,4096,0.1701
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,8192,0.1712
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,16,1,1,16384,0.1773
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,1,0.1494
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,4,0.1499
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,8,0.1512
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,16,0.1502
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,32,0.1497
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,64,0.1514
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,128,0.1518
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,256,0.1588
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,512,0.1601
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,1024,0.1615
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,2048,0.1747
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,32,1,1,4096,0.1782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,1,0.1573
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,4,0.1568
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,8,0.1572
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,16,0.1581
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,32,0.1565
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,64,0.1569
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,128,0.1591
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,256,0.1602
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,512,0.1622
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,1024,0.1667
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,2048,0.1838
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,64,1,1,4096,0.1880
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,1,0.1727
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,4,0.1737
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,8,0.1741
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,16,0.1730
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,32,0.1732
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,64,0.1732
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,128,0.1766
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,256,0.1782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,512,0.1825
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,1024,0.1915
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,128,1,1,2048,0.2160
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,1,0.2155
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,4,0.2153
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,8,0.2145
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,16,0.2152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,32,0.2161
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,64,0.2157
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,128,0.2182
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,256,0.2236
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,512,0.2325
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,256,1,1,1024,0.2523
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,1,0.2793
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,4,0.2788
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,8,0.2805
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,16,0.2788
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,32,0.2787
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,64,0.2797
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,128,0.2846
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,256,0.2956
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,512,1,1,512,0.3114
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1024,1,1,1,0.4476
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1024,1,1,4,0.4481
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1024,1,1,8,0.4489
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1024,1,1,16,0.4480
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1024,1,1,32,0.4493
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1024,1,1,64,0.4497
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1024,1,1,128,0.4607
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,128,1024,1,1,256,0.4784
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,1,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,4,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,8,0.1332
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,16,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,32,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,64,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,128,0.1338
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,256,0.1361
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,512,0.1360
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,1024,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,2048,0.1395
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,4096,0.1467
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,8192,0.1499
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1,1,1,16384,0.1500
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,1,0.1359
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,4,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,8,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,16,0.1374
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,32,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,64,0.1368
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,128,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,256,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,512,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,1024,0.1386
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,2048,0.1470
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,4096,0.1504
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,8192,0.1512
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,2,1,1,16384,0.1526
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,1,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,4,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,8,0.1361
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,16,0.1365
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,32,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,64,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,128,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,256,0.1387
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,512,0.1399
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,1024,0.1403
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,2048,0.1498
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,4096,0.1516
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,8192,0.1548
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,4,1,1,16384,0.1572
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,1,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,4,0.1360
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,8,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,16,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,32,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,64,0.1358
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,128,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,256,0.1401
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,512,0.1400
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,1024,0.1420
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,2048,0.1521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,4096,0.1539
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,8192,0.1554
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,8,1,1,16384,0.1596
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,1,0.1258
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,4,0.1248
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,8,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,16,0.1270
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,32,0.1255
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,64,0.1236
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,128,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,256,0.1314
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,512,0.1323
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,1024,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,2048,0.1481
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,4096,0.1520
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,8192,0.1538
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,16,1,1,16384,0.1597
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,1,0.1268
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,4,0.1272
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,8,0.1291
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,16,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,32,0.1272
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,64,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,128,0.1292
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,256,0.1340
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,512,0.1381
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,1024,0.1405
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,2048,0.1524
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,32,1,1,4096,0.1565
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,1,0.1340
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,4,0.1330
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,8,0.1337
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,16,0.1340
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,32,0.1335
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,64,0.1335
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,128,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,256,0.1441
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,512,0.1463
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,1024,0.1507
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,2048,0.1680
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,64,1,1,4096,0.1715
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,1,0.1414
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,4,0.1415
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,8,0.1416
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,16,0.1424
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,32,0.1412
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,64,0.1420
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,128,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,256,0.1484
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,512,0.1521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,1024,0.1615
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,128,1,1,2048,0.1854
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,1,0.1719
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,4,0.1706
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,8,0.1697
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,16,0.1724
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,32,0.1717
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,64,0.1704
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,128,0.1752
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,256,0.1792
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,512,0.1887
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,256,1,1,1024,0.2081
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,1,0.2177
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,4,0.2184
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,8,0.2196
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,16,0.2181
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,32,0.2181
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,64,0.2196
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,128,0.2288
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,256,0.2386
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,512,1,1,512,0.2562
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1024,1,1,1,0.3167
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1024,1,1,4,0.3175
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1024,1,1,8,0.3164
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1024,1,1,16,0.3163
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1024,1,1,32,0.3170
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1024,1,1,64,0.3193
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1024,1,1,128,0.3342
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,64,1024,1,1,256,0.3495
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1,1,1,1,0.0969
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1,1,1,4,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1,1,1,8,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1,1,1,16,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1,1,1,32,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1,1,1,64,0.1030
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1,1,1,128,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,2,1,1,1,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,2,1,1,4,0.1073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,2,1,1,8,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,2,1,1,16,0.1062
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,2,1,1,32,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,2,1,1,64,0.1073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,2,1,1,128,0.1079
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,4,1,1,1,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,4,1,1,4,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,4,1,1,8,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,4,1,1,16,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,4,1,1,32,0.1102
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,4,1,1,64,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,4,1,1,128,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,8,1,1,1,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,8,1,1,4,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,8,1,1,8,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,8,1,1,16,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,8,1,1,32,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,8,1,1,64,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,8,1,1,128,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,16,1,1,1,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,16,1,1,4,0.1094
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,16,1,1,8,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,16,1,1,16,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,16,1,1,32,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,16,1,1,64,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,16,1,1,128,0.1101
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,32,1,1,1,0.1053
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,32,1,1,4,0.1053
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,32,1,1,8,0.1052
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,32,1,1,16,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,32,1,1,32,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,32,1,1,64,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,32,1,1,128,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,64,1,1,1,0.1072
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,64,1,1,4,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,64,1,1,8,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,64,1,1,16,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,64,1,1,32,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,64,1,1,64,0.1095
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,64,1,1,128,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,128,1,1,1,0.1154
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,128,1,1,4,0.1171
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,128,1,1,8,0.1163
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,128,1,1,16,0.1140
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,128,1,1,32,0.1169
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,128,1,1,64,0.1178
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,128,1,1,128,0.1232
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,256,1,1,1,0.1438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,256,1,1,4,0.1445
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,256,1,1,8,0.1440
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,256,1,1,16,0.1426
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,256,1,1,32,0.1441
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,256,1,1,64,0.1444
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,256,1,1,128,0.1521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,512,1,1,1,0.1845
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,512,1,1,4,0.1846
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,512,1,1,8,0.1834
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,512,1,1,16,0.1831
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,512,1,1,32,0.1851
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,512,1,1,64,0.1844
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,512,1,1,128,0.1969
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1024,1,1,1,0.2570
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1024,1,1,4,0.2573
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1024,1,1,8,0.2575
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1024,1,1,16,0.2564
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1024,1,1,32,0.2585
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1024,1,1,64,0.2581
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,32,1024,1,1,128,0.2767
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1,1,1,1,0.0888
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1,1,1,4,0.0888
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1,1,1,8,0.0952
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1,1,1,16,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1,1,1,32,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1,1,1,64,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1,1,1,128,0.0965
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,2,1,1,1,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,2,1,1,4,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,2,1,1,8,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,2,1,1,16,0.0996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,2,1,1,32,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,2,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,2,1,1,128,0.1012
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,4,1,1,1,0.1019
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,4,1,1,4,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,4,1,1,8,0.1010
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,4,1,1,16,0.1013
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,4,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,4,1,1,64,0.1017
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,4,1,1,128,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,8,1,1,1,0.1012
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,8,1,1,4,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,8,1,1,8,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,8,1,1,16,0.1014
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,8,1,1,32,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,8,1,1,64,0.1021
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,8,1,1,128,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,16,1,1,1,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,16,1,1,4,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,16,1,1,8,0.1011
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,16,1,1,16,0.1017
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,16,1,1,32,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,16,1,1,64,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,16,1,1,128,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,32,1,1,1,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,32,1,1,4,0.1031
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,32,1,1,8,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,32,1,1,16,0.1032
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,32,1,1,32,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,32,1,1,64,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,32,1,1,128,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,64,1,1,1,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,64,1,1,4,0.1025
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,64,1,1,8,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,64,1,1,16,0.1019
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,64,1,1,32,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,64,1,1,64,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,64,1,1,128,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,128,1,1,1,0.1031
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,128,1,1,4,0.1053
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,128,1,1,8,0.1036
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,128,1,1,16,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,128,1,1,32,0.1063
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,128,1,1,64,0.1067
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,128,1,1,128,0.1078
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,256,1,1,1,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,256,1,1,4,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,256,1,1,8,0.1238
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,256,1,1,16,0.1277
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,256,1,1,32,0.1273
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,256,1,1,64,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,256,1,1,128,0.1359
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,512,1,1,1,0.1645
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,512,1,1,4,0.1645
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,512,1,1,8,0.1643
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,512,1,1,16,0.1642
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,512,1,1,32,0.1663
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,512,1,1,64,0.1646
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,512,1,1,128,0.1763
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1024,1,1,1,0.2260
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1024,1,1,4,0.2267
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1024,1,1,8,0.2267
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1024,1,1,16,0.2271
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1024,1,1,32,0.2281
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1024,1,1,64,0.2275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,16,1024,1,1,128,0.2460
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1,1,1,1,0.0846
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1,1,1,4,0.0845
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1,1,1,8,0.0907
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1,1,1,16,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1,1,1,32,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1,1,1,64,0.0907
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1,1,1,128,0.0926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,2,1,1,1,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,2,1,1,4,0.0960
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,2,1,1,8,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,2,1,1,16,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,2,1,1,32,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,2,1,1,64,0.0958
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,2,1,1,128,0.0975
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,4,1,1,1,0.0970
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,4,1,1,4,0.0980
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,4,1,1,8,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,4,1,1,16,0.0972
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,4,1,1,32,0.0994
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,4,1,1,64,0.0972
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,4,1,1,128,0.0974
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,8,1,1,1,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,8,1,1,4,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,8,1,1,8,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,8,1,1,16,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,8,1,1,32,0.0969
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,8,1,1,64,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,8,1,1,128,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,16,1,1,1,0.0989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,16,1,1,4,0.0989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,16,1,1,8,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,16,1,1,16,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,16,1,1,32,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,16,1,1,64,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,16,1,1,128,0.1010
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,32,1,1,1,0.0989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,32,1,1,4,0.0995
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,32,1,1,8,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,32,1,1,16,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,32,1,1,32,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,32,1,1,64,0.0996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,32,1,1,128,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,64,1,1,1,0.1003
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,64,1,1,4,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,64,1,1,8,0.1013
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,64,1,1,16,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,64,1,1,32,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,64,1,1,64,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,64,1,1,128,0.1019
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,128,1,1,1,0.1026
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,128,1,1,4,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,128,1,1,8,0.1019
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,128,1,1,16,0.1010
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,128,1,1,32,0.1030
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,128,1,1,64,0.1025
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,128,1,1,128,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,256,1,1,1,0.1152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,256,1,1,4,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,256,1,1,8,0.1155
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,256,1,1,16,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,256,1,1,32,0.1160
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,256,1,1,64,0.1194
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,256,1,1,128,0.1226
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,512,1,1,1,0.1514
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,512,1,1,4,0.1518
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,512,1,1,8,0.1517
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,512,1,1,16,0.1520
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,512,1,1,32,0.1537
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,512,1,1,64,0.1567
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,512,1,1,128,0.1684
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1024,1,1,1,0.2079
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1024,1,1,4,0.2094
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1024,1,1,8,0.2086
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1024,1,1,16,0.2091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1024,1,1,32,0.2096
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1024,1,1,64,0.2094
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,deepseek-ai/DeepSeek-V3.2,DeepseekV32ForCausalLM,float16,fp8,fp8_block,8,1024,1,1,128,0.2294
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,1,0.1249
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,4,0.1236
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,8,0.1244
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,16,0.1251
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,32,0.1247
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,64,0.1251
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,128,0.1283
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,256,0.1296
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,512,0.1300
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,1024,0.1309
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,2048,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,4096,0.1402
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,8192,0.1416
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1,1,1,16384,0.1446
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,1,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,4,0.1288
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,8,0.1287
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,16,0.1287
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,32,0.1289
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,64,0.1281
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,128,0.1315
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,256,0.1340
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,512,0.1335
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,1024,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,2048,0.1401
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,4096,0.1442
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,8192,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,2,1,1,16384,0.1481
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,1,0.1304
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,4,0.1303
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,8,0.1305
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,16,0.1312
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,32,0.1297
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,64,0.1309
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,128,0.1338
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,256,0.1358
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,512,0.1359
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,1024,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,2048,0.1469
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,4096,0.1497
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,8192,0.1511
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,4,1,1,16384,0.1543
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,1,0.1313
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,4,0.1318
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,8,0.1307
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,16,0.1318
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,32,0.1315
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,64,0.1313
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,128,0.1337
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,256,0.1367
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,512,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,1024,0.1426
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,2048,0.1526
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,4096,0.1560
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,8192,0.1575
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,8,1,1,16384,0.1617
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,1,0.1323
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,4,0.1333
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,8,0.1329
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,16,0.1326
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,32,0.1328
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,64,0.1328
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,128,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,256,0.1393
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,512,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,1024,0.1505
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,2048,0.1608
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,4096,0.1649
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,8192,0.1669
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,16,1,1,16384,0.1726
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,1,0.1381
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,4,0.1389
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,8,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,16,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,32,0.1378
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,64,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,128,0.1407
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,256,0.1493
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,512,0.1530
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,1024,0.1588
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,2048,0.1719
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,32,1,1,4096,0.1759
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,1,0.1451
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,4,0.1451
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,8,0.1456
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,16,0.1448
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,32,0.1453
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,64,0.1453
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,128,0.1484
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,256,0.1606
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,512,0.1644
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,1024,0.1723
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,2048,0.1959
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,64,1,1,4096,0.2013
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,1,0.1643
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,4,0.1660
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,8,0.1655
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,16,0.1657
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,32,0.1645
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,64,0.1666
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,128,0.1718
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,256,0.1756
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,512,0.1861
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,1024,0.2033
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,128,1,1,2048,0.2400
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,1,0.2013
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,4,0.2008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,8,0.2005
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,16,0.1999
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,32,0.2014
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,64,0.1999
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,128,0.2092
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,256,0.2191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,512,0.2405
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,256,1,1,1024,0.2751
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,1,0.2716
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,4,0.2729
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,8,0.2719
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,16,0.2720
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,32,0.2715
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,64,0.2732
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,128,0.2896
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,256,0.3109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,512,1,1,512,0.3479
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1024,1,1,1,0.4252
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1024,1,1,4,0.4274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1024,1,1,8,0.4269
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1024,1,1,16,0.4275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1024,1,1,32,0.4283
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1024,1,1,64,0.4306
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1024,1,1,128,0.4628
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,64,1024,1,1,256,0.5003
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,1,0.0996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,4,0.0996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,8,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,16,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,32,0.1012
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,64,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,128,0.1045
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,256,0.1067
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,512,0.1078
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,1024,0.1094
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,2048,0.1161
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,4096,0.1186
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,8192,0.1189
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1,1,1,16384,0.1209
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,1,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,4,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,8,0.1041
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,16,0.1046
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,32,0.1045
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,64,0.1046
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,128,0.1071
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,256,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,512,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,1024,0.1109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,2048,0.1177
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,4096,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,8192,0.1253
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,2,1,1,16384,0.1275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,1,0.1065
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,4,0.1061
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,8,0.1064
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,16,0.1067
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,32,0.1064
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,64,0.1067
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,128,0.1097
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,256,0.1114
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,512,0.1155
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,1024,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,2048,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,4096,0.1264
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,8192,0.1281
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,4,1,1,16384,0.1305
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,1,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,4,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,8,0.1084
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,16,0.1092
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,32,0.1088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,64,0.1092
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,128,0.1118
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,256,0.1173
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,512,0.1185
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,1024,0.1231
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,2048,0.1309
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,4096,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,8192,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,8,1,1,16384,0.1407
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,1,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,4,0.1102
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,8,0.1105
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,16,0.1106
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,32,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,64,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,128,0.1141
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,256,0.1194
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,512,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,1024,0.1292
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,2048,0.1376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,4096,0.1405
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,8192,0.1438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,16,1,1,16384,0.1502
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,1,0.1132
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,4,0.1132
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,8,0.1141
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,16,0.1137
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,32,0.1145
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,64,0.1141
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,128,0.1170
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,256,0.1266
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,512,0.1313
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,1024,0.1346
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,2048,0.1481
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,32,1,1,4096,0.1526
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,1,0.1275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,4,0.1276
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,8,0.1290
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,16,0.1276
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,32,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,64,0.1276
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,128,0.1333
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,256,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,512,0.1485
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,1024,0.1448
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,2048,0.1680
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,64,1,1,4096,0.1726
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,1,0.1291
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,4,0.1295
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,8,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,16,0.1295
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,32,0.1289
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,64,0.1294
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,128,0.1345
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,256,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,512,0.1469
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,1024,0.1636
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,128,1,1,2048,0.2012
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,1,0.1485
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,4,0.1484
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,8,0.1487
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,16,0.1485
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,32,0.1491
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,64,0.1492
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,128,0.1594
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,256,0.1677
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,512,0.1866
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,256,1,1,1024,0.2250
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,1,0.1981
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,4,0.1977
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,8,0.1983
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,16,0.1977
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,32,0.1971
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,64,0.1977
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,128,0.2149
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,256,0.2324
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,512,1,1,512,0.2658
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1024,1,1,1,0.3035
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1024,1,1,4,0.3042
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1024,1,1,8,0.3035
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1024,1,1,16,0.3035
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1024,1,1,32,0.3028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1024,1,1,64,0.3038
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1024,1,1,128,0.3354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,32,1024,1,1,256,0.3706
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,1,0.0825
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,4,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,8,0.0823
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,16,0.0862
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,32,0.0836
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,64,0.0827
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,128,0.0865
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,256,0.0889
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,512,0.0900
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,1024,0.0900
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,2048,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,4096,0.1011
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,8192,0.1033
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1,1,1,16384,0.1057
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,1,0.0904
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,4,0.0910
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,8,0.0916
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,16,0.0899
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,32,0.0907
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,64,0.0918
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,128,0.0934
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,256,0.0964
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,512,0.0963
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,1024,0.1010
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,2048,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,4096,0.1128
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,8192,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,2,1,1,16384,0.1189
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,1,0.0937
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,4,0.0928
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,8,0.0924
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,16,0.0942
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,32,0.0924
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,64,0.0936
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,128,0.0976
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,256,0.0996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,512,0.1025
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,1024,0.1040
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,2048,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,4096,0.1216
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,8192,0.1221
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,4,1,1,16384,0.1279
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,1,0.0960
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,4,0.0954
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,8,0.0958
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,16,0.0952
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,32,0.0979
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,64,0.0992
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,128,0.1043
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,256,0.1072
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,512,0.1084
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,1024,0.1141
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,2048,0.1258
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,4096,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,8192,0.1313
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,8,1,1,16384,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,1,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,4,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,8,0.0973
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,16,0.0980
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,32,0.0973
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,64,0.0990
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,128,0.1015
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,256,0.1063
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,512,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,1024,0.1194
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,2048,0.1311
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,4096,0.1340
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,8192,0.1381
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,16,1,1,16384,0.1428
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,1,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,4,0.0982
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,8,0.0994
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,16,0.0989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,32,0.1005
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,128,0.1054
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,256,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,512,0.1222
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,1024,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,2048,0.1405
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,32,1,1,4096,0.1446
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,1,0.1108
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,4,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,8,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,16,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,32,0.1092
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,64,0.1121
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,128,0.1178
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,256,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,512,0.1302
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,1024,0.1385
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,2048,0.1602
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,64,1,1,4096,0.1659
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,1,0.1182
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,4,0.1174
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,8,0.1192
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,16,0.1181
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,32,0.1174
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,64,0.1204
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,128,0.1266
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,256,0.1305
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,512,0.1392
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,1024,0.1558
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,128,1,1,2048,0.1940
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,1,0.1350
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,4,0.1365
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,8,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,16,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,32,0.1369
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,64,0.1368
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,128,0.1472
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,256,0.1555
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,512,0.1736
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,256,1,1,1024,0.2073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,1,0.1695
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,4,0.1686
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,8,0.1682
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,16,0.1678
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,32,0.1677
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,64,0.1694
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,128,0.1862
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,256,0.2045
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,512,1,1,512,0.2374
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1024,1,1,1,0.2453
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1024,1,1,4,0.2441
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1024,1,1,8,0.2446
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1024,1,1,16,0.2443
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1024,1,1,32,0.2435
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1024,1,1,64,0.2449
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1024,1,1,128,0.2758
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,16,1024,1,1,256,0.3069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,1,0.0759
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,4,0.0763
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,8,0.0760
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,16,0.0792
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,32,0.0760
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,64,0.0761
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,128,0.0762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,256,0.0782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,512,0.0783
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,1024,0.0787
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,2048,0.0829
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,4096,0.0896
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,8192,0.0905
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1,1,1,16384,0.0925
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,1,0.0822
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,4,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,8,0.0825
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,16,0.0803
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,32,0.0816
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,64,0.0825
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,128,0.0842
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,256,0.0835
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,512,0.0841
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,1024,0.0885
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,2048,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,4096,0.0971
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,8192,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,2,1,1,16384,0.1010
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,1,0.0847
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,4,0.0843
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,8,0.0843
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,16,0.0863
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,32,0.0845
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,64,0.0845
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,128,0.0868
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,256,0.0877
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,512,0.0863
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,1024,0.0868
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,2048,0.1004
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,4096,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,8192,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,4,1,1,16384,0.1103
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,1,0.0864
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,4,0.0853
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,8,0.0859
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,16,0.0857
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,32,0.0852
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,64,0.0865
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,128,0.0882
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,256,0.0870
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,512,0.0893
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,1024,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,2048,0.1054
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,4096,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,8192,0.1169
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,8,1,1,16384,0.1232
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,1,0.0823
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,4,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,8,0.0843
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,16,0.0835
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,32,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,64,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,128,0.0832
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,256,0.0873
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,512,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,1024,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,2048,0.1075
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,4096,0.1164
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,8192,0.1238
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,16,1,1,16384,0.1324
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,1,0.0866
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,4,0.0879
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,8,0.0867
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,16,0.0864
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,32,0.0865
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,64,0.0865
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,128,0.0886
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,256,0.0925
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,512,0.0945
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,1024,0.1134
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,2048,0.1318
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,32,1,1,4096,0.1366
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,1,0.0926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,4,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,8,0.0926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,16,0.0915
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,32,0.0918
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,64,0.0922
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,128,0.0975
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,256,0.1119
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,512,0.1192
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,1024,0.1306
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,2048,0.1532
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,64,1,1,4096,0.1580
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,1,0.0997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,4,0.0976
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,8,0.1017
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,16,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,32,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,64,0.1024
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,128,0.1112
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,256,0.1185
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,512,0.1315
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,1024,0.1499
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,128,1,1,2048,0.1872
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,1,0.1197
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,4,0.1203
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,8,0.1212
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,16,0.1220
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,32,0.1226
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,64,0.1250
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,128,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,256,0.1438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,512,0.1635
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,256,1,1,1024,0.1980
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,1,0.1518
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,4,0.1504
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,8,0.1498
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,16,0.1504
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,32,0.1506
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,64,0.1527
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,128,0.1696
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,256,0.1853
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,512,1,1,512,0.2194
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1024,1,1,1,0.2110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1024,1,1,4,0.2097
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1024,1,1,8,0.2094
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1024,1,1,16,0.2104
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1024,1,1,32,0.2108
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1024,1,1,64,0.2114
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1024,1,1,128,0.2412
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,float16,8,1024,1,1,256,0.2714
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,1,0.1191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,4,0.1184
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,8,0.1184
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,16,0.1194
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,32,0.1186
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,64,0.1189
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,128,0.1194
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,256,0.1215
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,512,0.1219
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,1024,0.1231
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,2048,0.1281
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,4096,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,8192,0.1333
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1,1,1,16384,0.1370
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,1,0.1209
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,4,0.1210
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,8,0.1214
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,16,0.1209
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,32,0.1213
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,64,0.1211
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,128,0.1219
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,256,0.1244
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,512,0.1240
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,1024,0.1240
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,2048,0.1299
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,4096,0.1350
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,8192,0.1359
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,2,1,1,16384,0.1384
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,1,0.1229
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,4,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,8,0.1236
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,16,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,32,0.1229
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,64,0.1237
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,128,0.1237
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,256,0.1260
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,512,0.1261
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,1024,0.1272
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,2048,0.1367
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,4096,0.1395
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,8192,0.1399
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,4,1,1,16384,0.1426
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,1,0.1223
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,4,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,8,0.1227
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,16,0.1231
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,32,0.1229
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,64,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,128,0.1238
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,256,0.1268
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,512,0.1261
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,1024,0.1284
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,2048,0.1389
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,4096,0.1417
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,8192,0.1437
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,8,1,1,16384,0.1469
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,1,0.1246
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,4,0.1249
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,8,0.1243
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,16,0.1237
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,32,0.1239
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,64,0.1238
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,128,0.1249
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,256,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,512,0.1297
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,1024,0.1354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,2048,0.1455
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,4096,0.1492
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,8192,0.1513
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,16,1,1,16384,0.1562
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,1,0.1291
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,4,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,8,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,16,0.1292
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,32,0.1285
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,64,0.1291
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,128,0.1298
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,256,0.1353
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,512,0.1381
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,1024,0.1416
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,2048,0.1519
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,32,1,1,4096,0.1564
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,1,0.1352
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,4,0.1348
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,8,0.1350
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,16,0.1346
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,32,0.1344
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,64,0.1352
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,128,0.1362
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,256,0.1447
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,512,0.1471
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,1024,0.1521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,2048,0.1680
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,64,1,1,4096,0.1729
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,1,0.1430
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,4,0.1433
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,8,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,16,0.1439
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,32,0.1434
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,64,0.1432
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,128,0.1462
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,256,0.1494
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,512,0.1540
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,1024,0.1623
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,128,1,1,2048,0.1858
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,1,0.1758
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,4,0.1762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,8,0.1763
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,16,0.1756
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,32,0.1754
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,64,0.1760
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,128,0.1806
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,256,0.1843
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,512,0.1931
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,256,1,1,1024,0.2132
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,1,0.2453
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,4,0.2452
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,8,0.2448
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,16,0.2447
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,32,0.2455
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,64,0.2455
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,128,0.2535
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,256,0.2618
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,512,1,1,512,0.2822
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1024,1,1,1,0.3878
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1024,1,1,4,0.3899
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1024,1,1,8,0.3883
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1024,1,1,16,0.3887
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1024,1,1,32,0.3887
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1024,1,1,64,0.3907
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1024,1,1,128,0.4059
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,64,1024,1,1,256,0.4232
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,1,0.0954
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,4,0.0950
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,8,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,16,0.0961
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,32,0.0965
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,64,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,128,0.0973
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,256,0.0997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,512,0.0995
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,1024,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,2048,0.1035
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,4096,0.1099
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,8192,0.1107
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1,1,1,16384,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,1,0.0973
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,4,0.0973
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,8,0.0973
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,16,0.0969
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,32,0.0972
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,64,0.0969
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,128,0.0981
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,256,0.0994
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,512,0.0997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,1024,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,2048,0.1073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,4096,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,8192,0.1140
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,2,1,1,16384,0.1166
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,1,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,4,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,8,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,16,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,32,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,64,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,128,0.1000
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,256,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,512,0.1036
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,1024,0.1034
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,2048,0.1128
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,4096,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,8192,0.1158
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,4,1,1,16384,0.1194
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,1,0.1002
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,4,0.1005
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,8,0.0998
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,16,0.0997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,32,0.0997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,64,0.0995
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,128,0.1004
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,256,0.1039
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,512,0.1043
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,1024,0.1075
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,2048,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,4096,0.1184
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,8192,0.1196
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,8,1,1,16384,0.1241
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,1,0.1026
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,4,0.1026
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,8,0.1026
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,16,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,32,0.1025
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,64,0.1025
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,128,0.1026
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,256,0.1071
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,512,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,1024,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,2048,0.1227
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,4096,0.1258
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,8192,0.1278
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,16,1,1,16384,0.1340
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,1,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,4,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,8,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,16,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,32,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,64,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,128,0.1064
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,256,0.1128
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,512,0.1160
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,1024,0.1173
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,2048,0.1289
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,32,1,1,4096,0.1332
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,1,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,4,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,8,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,16,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,32,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,64,0.1072
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,128,0.1088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,256,0.1192
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,512,0.1217
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,1024,0.1255
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,2048,0.1406
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,64,1,1,4096,0.1459
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,1,0.1154
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,4,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,8,0.1157
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,16,0.1152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,32,0.1157
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,64,0.1161
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,128,0.1197
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,256,0.1214
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,512,0.1262
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,1024,0.1353
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,128,1,1,2048,0.1572
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,1,0.1343
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,4,0.1354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,8,0.1344
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,16,0.1350
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,32,0.1353
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,64,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,128,0.1407
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,256,0.1463
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,512,0.1532
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,256,1,1,1024,0.1739
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,1,0.1776
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,4,0.1775
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,8,0.1780
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,16,0.1775
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,32,0.1774
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,64,0.1778
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,128,0.1887
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,256,0.1950
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,512,1,1,512,0.2135
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1024,1,1,1,0.2757
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1024,1,1,4,0.2750
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1024,1,1,8,0.2752
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1024,1,1,16,0.2756
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1024,1,1,32,0.2746
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1024,1,1,64,0.2755
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1024,1,1,128,0.2941
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,32,1024,1,1,256,0.3068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,1,0.0787
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,4,0.0782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,8,0.0782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,16,0.0822
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,32,0.0790
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,64,0.0787
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,128,0.0802
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,256,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,512,0.0821
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,1024,0.0831
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,2048,0.0861
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,4096,0.0930
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,8192,0.0949
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1,1,1,16384,0.0974
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,1,0.0804
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,4,0.0802
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,8,0.0806
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,16,0.0802
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,32,0.0809
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,64,0.0804
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,128,0.0810
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,256,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,512,0.0827
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,1024,0.0828
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,2048,0.0927
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,4096,0.0942
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,8192,0.0949
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,2,1,1,16384,0.0985
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,1,0.0802
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,4,0.0804
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,8,0.0804
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,16,0.0803
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,32,0.0804
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,64,0.0803
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,128,0.0823
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,256,0.0822
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,512,0.0837
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,1024,0.0857
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,2048,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,4096,0.0984
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,8192,0.1011
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,4,1,1,16384,0.1056
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,1,0.0834
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,4,0.0836
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,8,0.0828
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,16,0.0833
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,32,0.0843
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,64,0.0839
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,128,0.0855
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,256,0.0866
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,512,0.0888
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,1024,0.0896
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,2048,0.1003
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,4096,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,8192,0.1065
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,8,1,1,16384,0.1112
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,1,0.0825
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,4,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,8,0.0820
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,16,0.0822
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,32,0.0823
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,64,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,128,0.0833
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,256,0.0873
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,512,0.0885
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,1024,0.0927
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,2048,0.1060
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,4096,0.1093
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,8192,0.1123
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,16,1,1,16384,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,1,0.0844
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,4,0.0845
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,8,0.0849
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,16,0.0854
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,32,0.0858
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,64,0.0855
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,128,0.0874
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,256,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,512,0.0945
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,1024,0.0978
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,2048,0.1097
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,32,1,1,4096,0.1132
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,1,0.0907
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,4,0.0905
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,8,0.0891
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,16,0.0897
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,32,0.0904
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,64,0.0915
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,128,0.0937
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,256,0.0990
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,512,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,1024,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,2048,0.1218
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,64,1,1,4096,0.1277
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,1,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,4,0.0980
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,8,0.0978
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,16,0.0981
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,32,0.0978
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,64,0.0989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,128,0.1011
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,256,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,512,0.1082
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,1024,0.1159
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,128,1,1,2048,0.1378
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,1,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,4,0.1132
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,8,0.1128
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,16,0.1121
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,32,0.1132
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,64,0.1134
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,128,0.1191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,256,0.1232
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,512,0.1306
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,256,1,1,1024,0.1483
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,1,0.1420
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,4,0.1422
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,8,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,16,0.1427
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,32,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,64,0.1432
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,128,0.1521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,256,0.1588
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,512,1,1,512,0.1764
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1024,1,1,1,0.2103
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1024,1,1,4,0.2111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1024,1,1,8,0.2110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1024,1,1,16,0.2110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1024,1,1,32,0.2101
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1024,1,1,64,0.2107
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1024,1,1,128,0.2288
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,16,1024,1,1,256,0.2423
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,1,0.0716
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,4,0.0720
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,8,0.0718
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,16,0.0757
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,32,0.0720
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,64,0.0712
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,128,0.0722
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,256,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,512,0.0741
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,1024,0.0746
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,2048,0.0785
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,4096,0.0844
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,8192,0.0863
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1,1,1,16384,0.0886
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,1,0.0720
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,4,0.0722
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,8,0.0721
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,16,0.0721
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,32,0.0721
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,64,0.0722
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,128,0.0732
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,256,0.0740
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,512,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,1024,0.0765
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,2048,0.0843
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,4096,0.0879
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,8192,0.0886
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,2,1,1,16384,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,1,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,4,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,8,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,16,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,32,0.0741
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,64,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,128,0.0757
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,256,0.0761
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,512,0.0762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,1024,0.0762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,2048,0.0865
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,4096,0.0901
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,8192,0.0904
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,4,1,1,16384,0.0937
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,1,0.0743
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,4,0.0741
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,8,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,16,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,32,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,64,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,128,0.0750
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,256,0.0762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,512,0.0765
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,1024,0.0763
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,2048,0.0887
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,4096,0.0927
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,8192,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,8,1,1,16384,0.1000
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,1,0.0717
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,4,0.0720
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,8,0.0711
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,16,0.0709
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,32,0.0713
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,64,0.0712
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,128,0.0721
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,256,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,512,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,1024,0.0756
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,2048,0.0885
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,4096,0.0926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,8192,0.0971
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,16,1,1,16384,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,1,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,4,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,8,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,16,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,32,0.0741
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,64,0.0742
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,128,0.0760
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,256,0.0778
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,512,0.0783
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,1024,0.0826
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,2048,0.0971
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,32,1,1,4096,0.1033
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,1,0.0798
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,4,0.0800
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,8,0.0795
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,16,0.0793
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,32,0.0795
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,64,0.0799
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,128,0.0804
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,256,0.0855
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,512,0.0903
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,1024,0.0976
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,2048,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,64,1,1,4096,0.1205
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,1,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,4,0.0824
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,8,0.0845
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,16,0.0828
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,32,0.0844
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,64,0.0845
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,128,0.0866
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,256,0.0906
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,512,0.0996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,1024,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,128,1,1,2048,0.1311
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,1,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,4,0.0977
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,8,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,16,0.0985
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,32,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,64,0.0995
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,128,0.1061
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,256,0.1127
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,512,0.1211
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,256,1,1,1024,0.1386
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,1,0.1248
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,4,0.1240
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,8,0.1238
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,16,0.1242
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,32,0.1251
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,64,0.1256
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,128,0.1354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,256,0.1426
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,512,1,1,512,0.1590
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1024,1,1,1,0.1778
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1024,1,1,4,0.1771
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1024,1,1,8,0.1767
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1024,1,1,16,0.1772
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1024,1,1,32,0.1768
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1024,1,1,64,0.1775
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1024,1,1,128,0.1967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,float16,8,1024,1,1,256,0.2089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,1,0.1540
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,4,0.1545
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,8,0.1658
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,16,0.1650
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,32,0.1656
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,64,0.1647
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,128,0.1669
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,256,0.1717
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,512,0.1688
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,1024,0.1713
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,2048,0.1782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,4096,0.1801
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,8192,0.1808
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1,1,1,16384,0.1826
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,1,0.1758
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,4,0.1746
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,8,0.1754
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,16,0.1746
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,32,0.1762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,64,0.1747
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,128,0.1784
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,256,0.1778
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,512,0.1789
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,1024,0.1803
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,2048,0.1895
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,4096,0.1926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,8192,0.1935
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,2,1,1,16384,0.1955
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,1,0.1729
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,4,0.1728
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,8,0.1739
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,16,0.1734
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,32,0.1741
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,64,0.1732
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,128,0.1754
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,256,0.1776
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,512,0.1775
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,1024,0.1785
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,2048,0.1887
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,4096,0.1910
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,8192,0.1914
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,4,1,1,16384,0.1948
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,1,0.1487
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,4,0.1478
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,8,0.1478
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,16,0.1486
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,32,0.1496
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,64,0.1477
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,128,0.1520
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,256,0.1536
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,512,0.1553
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,1024,0.1602
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,2048,0.1703
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,4096,0.1729
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,8192,0.1752
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,8,1,1,16384,0.1782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,1,0.1498
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,4,0.1512
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,8,0.1502
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,16,0.1500
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,32,0.1503
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,64,0.1504
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,128,0.1525
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,256,0.1580
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,512,0.1609
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,1024,0.1684
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,2048,0.1797
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,4096,0.1815
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,8192,0.1849
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,16,1,1,16384,0.1907
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,1,0.1543
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,4,0.1539
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,8,0.1535
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,16,0.1535
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,32,0.1553
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,64,0.1525
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,128,0.1550
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,256,0.1652
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,512,0.1704
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,1024,0.1737
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,2048,0.1890
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,32,1,1,4096,0.1922
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,1,0.1555
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,4,0.1552
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,8,0.1551
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,16,0.1553
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,32,0.1564
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,64,0.1548
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,128,0.1592
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,256,0.1701
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,512,0.1753
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,1024,0.1832
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,2048,0.2063
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,64,1,1,4096,0.2126
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,1,0.1661
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,4,0.1657
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,8,0.1663
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,16,0.1670
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,32,0.1659
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,64,0.1665
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,128,0.1725
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,256,0.1754
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,512,0.1862
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,1024,0.2049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,128,1,1,2048,0.2415
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,1,0.1928
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,4,0.1919
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,8,0.1909
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,16,0.1920
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,32,0.1918
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,64,0.1914
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,128,0.2008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,256,0.2103
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,512,0.2302
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,256,1,1,1024,0.2682
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,1,0.2445
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,4,0.2449
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,8,0.2445
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,16,0.2454
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,32,0.2438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,64,0.2472
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,128,0.2640
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,256,0.2859
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,512,1,1,512,0.3199
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1024,1,1,1,0.3652
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1024,1,1,4,0.3646
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1024,1,1,8,0.3660
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1024,1,1,16,0.3650
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1024,1,1,32,0.3668
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1024,1,1,64,0.3706
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1024,1,1,128,0.4021
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,64,1024,1,1,256,0.4376
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,1,0.1231
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,4,0.1216
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,8,0.1314
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,16,0.1312
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,32,0.1304
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,64,0.1303
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,128,0.1327
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,256,0.1354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,512,0.1353
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,1024,0.1383
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,2048,0.1449
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,4096,0.1479
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,8192,0.1485
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1,1,1,16384,0.1503
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,1,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,4,0.1349
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,8,0.1351
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,16,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,32,0.1369
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,64,0.1338
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,128,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,256,0.1365
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,512,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,1024,0.1377
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,2048,0.1503
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,4096,0.1539
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,8192,0.1537
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,2,1,1,16384,0.1576
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,1,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,4,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,8,0.1347
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,16,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,32,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,64,0.1353
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,128,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,256,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,512,0.1411
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,1024,0.1463
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,2048,0.1579
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,4096,0.1602
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,8192,0.1602
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,4,1,1,16384,0.1663
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,1,0.1353
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,4,0.1344
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,8,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,16,0.1330
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,32,0.1346
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,64,0.1316
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,128,0.1364
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,256,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,512,0.1461
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,1024,0.1562
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,2048,0.1647
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,4096,0.1670
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,8192,0.1705
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,8,1,1,16384,0.1740
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,1,0.1253
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,4,0.1267
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,8,0.1261
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,16,0.1252
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,32,0.1248
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,64,0.1281
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,128,0.1316
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,256,0.1400
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,512,0.1413
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,1024,0.1512
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,2048,0.1602
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,4096,0.1644
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,8192,0.1654
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,16,1,1,16384,0.1723
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,1,0.1294
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,4,0.1318
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,8,0.1287
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,16,0.1315
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,32,0.1301
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,64,0.1331
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,128,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,256,0.1446
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,512,0.1516
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,1024,0.1548
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,2048,0.1693
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,32,1,1,4096,0.1733
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,1,0.1344
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,4,0.1351
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,8,0.1337
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,16,0.1353
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,32,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,64,0.1359
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,128,0.1389
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,256,0.1523
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,512,0.1553
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,1024,0.1658
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,2048,0.1869
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,64,1,1,4096,0.1923
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,1,0.1441
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,4,0.1446
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,8,0.1443
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,16,0.1445
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,32,0.1443
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,64,0.1463
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,128,0.1496
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,256,0.1555
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,512,0.1613
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,1024,0.1786
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,128,1,1,2048,0.2175
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,1,0.1596
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,4,0.1621
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,8,0.1600
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,16,0.1606
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,32,0.1604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,64,0.1616
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,128,0.1701
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,256,0.1816
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,512,0.1987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,256,1,1,1024,0.2371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,1,0.1990
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,4,0.1989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,8,0.1992
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,16,0.1994
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,32,0.1996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,64,0.1992
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,128,0.2172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,256,0.2352
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,512,1,1,512,0.2688
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1024,1,1,1,0.2788
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1024,1,1,4,0.2804
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1024,1,1,8,0.2786
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1024,1,1,16,0.2801
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1024,1,1,32,0.2794
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1024,1,1,64,0.2822
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1024,1,1,128,0.3138
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,32,1024,1,1,256,0.3505
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,1,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,4,0.0997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,8,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,16,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,32,0.1052
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,64,0.1057
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,128,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,256,0.1109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,512,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,1024,0.1096
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,2048,0.1191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,4096,0.1201
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,8192,0.1213
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1,1,1,16384,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,1,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,4,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,8,0.1123
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,16,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,32,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,64,0.1121
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,128,0.1134
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,256,0.1137
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,512,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,1024,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,2048,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,4096,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,8192,0.1272
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,2,1,1,16384,0.1291
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,1,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,4,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,8,0.1115
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,16,0.1134
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,32,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,64,0.1114
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,128,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,256,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,512,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,1024,0.1152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,2048,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,4096,0.1295
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,8192,0.1322
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,4,1,1,16384,0.1374
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,1,0.1152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,4,0.1149
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,8,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,16,0.1152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,32,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,64,0.1134
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,128,0.1173
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,256,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,512,0.1192
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,1024,0.1191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,2048,0.1320
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,4096,0.1356
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,8192,0.1398
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,8,1,1,16384,0.1468
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,1,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,4,0.1173
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,8,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,16,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,32,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,64,0.1173
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,128,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,256,0.1219
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,512,0.1197
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,1024,0.1235
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,2048,0.1478
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,4096,0.1516
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,8192,0.1550
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,16,1,1,16384,0.1645
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,1,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,4,0.1107
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,8,0.1088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,16,0.1109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,32,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,64,0.1095
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,128,0.1117
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,256,0.1171
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,512,0.1191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,1024,0.1302
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,2048,0.1557
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,32,1,1,4096,0.1607
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,1,0.1058
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,4,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,8,0.1056
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,16,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,32,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,64,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,128,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,256,0.1218
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,512,0.1342
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,1024,0.1480
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,2048,0.1713
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,64,1,1,4096,0.1761
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,1,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,4,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,8,0.1132
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,16,0.1146
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,32,0.1149
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,64,0.1174
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,128,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,256,0.1349
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,512,0.1483
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,1024,0.1656
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,128,1,1,2048,0.2035
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,1,0.1310
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,4,0.1323
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,8,0.1311
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,16,0.1311
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,32,0.1335
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,64,0.1350
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,128,0.1516
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,256,0.1640
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,512,0.1821
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,256,1,1,1024,0.2172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,1,0.1713
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,4,0.1706
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,8,0.1714
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,16,0.1714
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,32,0.1731
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,64,0.1737
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,128,0.1913
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,256,0.2073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,512,1,1,512,0.2408
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1024,1,1,1,0.2359
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1024,1,1,4,0.2371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1024,1,1,8,0.2363
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1024,1,1,16,0.2370
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1024,1,1,32,0.2358
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1024,1,1,64,0.2388
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1024,1,1,128,0.2698
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,16,1024,1,1,256,0.2996
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,1,0.0926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,4,0.0926
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,8,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,16,0.0989
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,32,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,64,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,128,0.0997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,256,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,512,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,1024,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,2048,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,4096,0.1134
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,8192,0.1143
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1,1,1,16384,0.1152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,1,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,4,0.1030
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,8,0.1033
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,16,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,32,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,64,0.1043
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,128,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,256,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,512,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,1024,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,2048,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,4096,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,8192,0.1181
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,2,1,1,16384,0.1191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,1,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,4,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,8,0.1048
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,16,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,32,0.1040
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,64,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,128,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,256,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,512,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,1024,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,2048,0.1195
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,4096,0.1202
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,8192,0.1225
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,4,1,1,16384,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,1,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,4,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,8,0.1055
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,16,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,32,0.1060
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,64,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,128,0.1085
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,256,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,512,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,1024,0.1109
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,2048,0.1216
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,4096,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,8192,0.1266
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,8,1,1,16384,0.1331
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,1,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,4,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,8,0.1073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,16,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,32,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,64,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,128,0.1075
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,256,0.1136
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,512,0.1140
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,1024,0.1138
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,2048,0.1294
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,4096,0.1350
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,8192,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,16,1,1,16384,0.1483
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,1,0.1084
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,4,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,8,0.1072
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,16,0.1087
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,32,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,64,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,128,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,256,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,512,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,1024,0.1197
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,2048,0.1424
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,32,1,1,4096,0.1545
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,1,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,4,0.1037
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,8,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,16,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,32,0.1033
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,64,0.1033
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,128,0.1054
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,256,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,512,0.1135
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,1024,0.1358
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,2048,0.1660
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,64,1,1,4096,0.1706
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,1,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,4,0.1071
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,8,0.1060
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,16,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,32,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,64,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,128,0.1113
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,256,0.1214
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,512,0.1420
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,1024,0.1609
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,128,1,1,2048,0.1986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,1,0.1205
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,4,0.1230
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,8,0.1215
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,16,0.1223
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,32,0.1236
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,64,0.1250
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,128,0.1392
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,256,0.1574
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,512,0.1733
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,256,1,1,1024,0.2069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,1,0.1493
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,4,0.1467
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,8,0.1492
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,16,0.1504
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,32,0.1543
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,64,0.1598
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,128,0.1783
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,256,0.1951
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,512,1,1,512,0.2290
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1024,1,1,1,0.2099
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1024,1,1,4,0.2135
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1024,1,1,8,0.2129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1024,1,1,16,0.2148
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1024,1,1,32,0.2138
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1024,1,1,64,0.2167
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1024,1,1,128,0.2468
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,float16,fp8_block,8,1024,1,1,256,0.2747
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,1,0.1486
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,4,0.1480
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,8,0.1582
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,16,0.1582
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,32,0.1586
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,64,0.1582
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,128,0.1595
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,256,0.1615
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,512,0.1622
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,1024,0.1623
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,2048,0.1696
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,4096,0.1712
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,8192,0.1734
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1,1,1,16384,0.1747
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,1,0.1615
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,4,0.1615
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,8,0.1624
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,16,0.1607
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,32,0.1614
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,64,0.1620
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,128,0.1616
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,256,0.1642
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,512,0.1661
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,1024,0.1647
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,2048,0.1736
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,4096,0.1762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,8192,0.1769
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,2,1,1,16384,0.1784
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,1,0.1626
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,4,0.1643
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,8,0.1627
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,16,0.1637
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,32,0.1631
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,64,0.1642
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,128,0.1652
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,256,0.1652
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,512,0.1665
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,1024,0.1684
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,2048,0.1758
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,4096,0.1782
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,8192,0.1809
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,4,1,1,16384,0.1830
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,1,0.1387
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,4,0.1393
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,8,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,16,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,32,0.1404
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,64,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,128,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,256,0.1428
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,512,0.1430
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,1024,0.1444
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,2048,0.1540
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,4096,0.1568
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,8192,0.1590
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,8,1,1,16384,0.1628
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,1,0.1406
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,4,0.1399
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,8,0.1403
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,16,0.1415
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,32,0.1398
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,64,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,128,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,256,0.1448
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,512,0.1458
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,1024,0.1510
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,2048,0.1617
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,4096,0.1647
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,8192,0.1665
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,16,1,1,16384,0.1731
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,1,0.1440
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,4,0.1439
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,8,0.1442
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,16,0.1432
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,32,0.1438
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,64,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,128,0.1446
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,256,0.1506
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,512,0.1535
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,1024,0.1550
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,2048,0.1676
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,32,1,1,4096,0.1714
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,1,0.1439
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,4,0.1442
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,8,0.1439
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,16,0.1451
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,32,0.1435
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,64,0.1428
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,128,0.1460
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,256,0.1549
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,512,0.1566
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,1024,0.1607
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,2048,0.1777
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,64,1,1,4096,0.1817
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,1,0.1528
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,4,0.1524
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,8,0.1536
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,16,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,32,0.1529
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,64,0.1538
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,128,0.1573
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,256,0.1588
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,512,0.1637
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,1024,0.1728
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,128,1,1,2048,0.1966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,1,0.1776
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,4,0.1768
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,8,0.1762
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,16,0.1776
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,32,0.1771
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,64,0.1759
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,128,0.1813
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,256,0.1852
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,512,0.1966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,256,1,1,1024,0.2148
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,1,0.2247
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,4,0.2253
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,8,0.2258
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,16,0.2255
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,32,0.2254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,64,0.2258
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,128,0.2357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,256,0.2448
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,512,1,1,512,0.2635
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1024,1,1,1,0.3360
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1024,1,1,4,0.3352
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1024,1,1,8,0.3340
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1024,1,1,16,0.3355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1024,1,1,32,0.3363
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1024,1,1,64,0.3384
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1024,1,1,128,0.3525
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,64,1024,1,1,256,0.3678
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,1,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,4,0.1166
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,8,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,16,0.1252
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,32,0.1253
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,64,0.1238
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,128,0.1260
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,256,0.1280
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,512,0.1256
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,1024,0.1297
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,2048,0.1395
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,4096,0.1418
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,8192,0.1434
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1,1,1,16384,0.1442
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,1,0.1298
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,4,0.1283
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,8,0.1301
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,16,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,32,0.1287
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,64,0.1296
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,128,0.1312
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,256,0.1295
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,512,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,1024,0.1295
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,2048,0.1396
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,4096,0.1425
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,8192,0.1439
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,2,1,1,16384,0.1463
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,1,0.1275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,4,0.1261
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,8,0.1279
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,16,0.1275
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,32,0.1252
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,64,0.1272
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,128,0.1276
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,256,0.1312
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,512,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,1024,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,2048,0.1419
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,4096,0.1446
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,8192,0.1458
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,4,1,1,16384,0.1500
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,1,0.1273
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,4,0.1279
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,8,0.1282
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,16,0.1274
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,32,0.1255
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,64,0.1253
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,128,0.1278
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,256,0.1298
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,512,0.1294
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,1024,0.1354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,2048,0.1483
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,4096,0.1480
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,8192,0.1519
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,8,1,1,16384,0.1586
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,1,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,4,0.1167
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,8,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,16,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,32,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,64,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,128,0.1170
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,256,0.1211
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,512,0.1236
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,1024,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,2048,0.1426
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,4096,0.1458
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,8192,0.1500
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,16,1,1,16384,0.1553
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,1,0.1202
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,4,0.1209
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,8,0.1206
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,16,0.1217
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,32,0.1205
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,64,0.1190
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,128,0.1229
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,256,0.1299
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,512,0.1347
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,1024,0.1371
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,2048,0.1475
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,32,1,1,4096,0.1516
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,1,0.1231
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,4,0.1210
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,8,0.1221
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,16,0.1215
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,32,0.1214
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,64,0.1228
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,128,0.1246
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,256,0.1357
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,512,0.1395
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,1024,0.1410
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,2048,0.1587
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,64,1,1,4096,0.1621
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,1,0.1296
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,4,0.1287
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,8,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,16,0.1304
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,32,0.1281
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,64,0.1307
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,128,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,256,0.1362
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,512,0.1409
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,1024,0.1494
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,128,1,1,2048,0.1721
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,1,0.1459
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,4,0.1448
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,8,0.1447
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,16,0.1457
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,32,0.1456
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,64,0.1461
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,128,0.1524
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,256,0.1559
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,512,0.1662
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,256,1,1,1024,0.1833
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,1,0.1788
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,4,0.1791
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,8,0.1793
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,16,0.1788
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,32,0.1791
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,64,0.1796
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,128,0.1909
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,256,0.1974
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,512,1,1,512,0.2165
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1024,1,1,1,0.2519
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1024,1,1,4,0.2514
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1024,1,1,8,0.2521
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1024,1,1,16,0.2530
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1024,1,1,32,0.2520
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1024,1,1,64,0.2541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1024,1,1,128,0.2725
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,32,1024,1,1,256,0.2867
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,1,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,4,0.0966
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,8,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,16,0.1011
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,32,0.1022
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,64,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,128,0.1039
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,256,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,512,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,1024,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,2048,0.1135
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,4096,0.1170
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,8192,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1,1,1,16384,0.1192
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,1,0.1066
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,4,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,8,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,16,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,32,0.1054
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,64,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,128,0.1072
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,256,0.1088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,512,0.1088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,1024,0.1076
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,2048,0.1177
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,4096,0.1193
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,8192,0.1191
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,2,1,1,16384,0.1218
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,1,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,4,0.1068
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,8,0.1057
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,16,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,32,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,64,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,128,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,256,0.1091
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,512,0.1082
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,1024,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,2048,0.1193
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,4096,0.1217
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,8192,0.1232
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,4,1,1,16384,0.1254
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,1,0.1073
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,4,0.1088
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,8,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,16,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,32,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,64,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,128,0.1093
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,256,0.1096
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,512,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,1024,0.1106
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,2048,0.1233
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,4096,0.1241
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,8192,0.1265
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,8,1,1,16384,0.1328
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,1,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,4,0.1099
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,8,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,16,0.1096
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,32,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,64,0.1089
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,128,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,256,0.1121
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,512,0.1110
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,1024,0.1132
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,2048,0.1267
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,4096,0.1276
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,8192,0.1375
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,16,1,1,16384,0.1468
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,1,0.1025
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,4,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,8,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,16,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,32,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,64,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,128,0.1032
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,256,0.1051
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,512,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,1024,0.1112
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,2048,0.1293
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,32,1,1,4096,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,1,0.1013
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,4,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,8,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,16,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,32,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,64,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,128,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,256,0.1049
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,512,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,1024,0.1221
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,2048,0.1317
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,64,1,1,4096,0.1354
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,1,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,4,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,8,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,16,0.0998
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,32,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,64,0.1000
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,128,0.1021
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,256,0.1054
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,512,0.1149
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,1024,0.1239
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,128,1,1,2048,0.1462
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,1,0.1121
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,4,0.1125
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,8,0.1119
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,16,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,32,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,64,0.1143
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,128,0.1211
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,256,0.1282
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,512,0.1362
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,256,1,1,1024,0.1539
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,1,0.1435
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,4,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,8,0.1428
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,16,0.1436
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,32,0.1441
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,64,0.1439
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,128,0.1541
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,256,0.1604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,512,1,1,512,0.1785
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1024,1,1,1,0.2005
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1024,1,1,4,0.2004
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1024,1,1,8,0.2001
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1024,1,1,16,0.2001
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1024,1,1,32,0.1997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1024,1,1,64,0.2007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1024,1,1,128,0.2186
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,16,1024,1,1,256,0.2320
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,1,0.0905
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,4,0.0904
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,8,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,16,0.0948
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,32,0.0955
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,64,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,128,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,256,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,512,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,1024,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,2048,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,4096,0.1105
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,8192,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1,1,1,16384,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,1,0.0997
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,4,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,8,0.1013
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,16,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,32,0.0992
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,128,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,256,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,512,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,1024,0.1016
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,2048,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,4096,0.1131
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,8192,0.1130
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,2,1,1,16384,0.1156
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,1,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,4,0.1002
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,8,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,16,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,32,0.1006
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,64,0.0987
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,128,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,256,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,512,0.1015
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,1024,0.1016
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,2048,0.1129
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,4096,0.1150
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,8192,0.1152
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,4,1,1,16384,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,1,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,4,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,8,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,16,0.0992
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,32,0.0992
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,64,0.1000
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,128,0.1014
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,256,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,512,0.1015
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,1024,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,2048,0.1145
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,4096,0.1151
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,8192,0.1172
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,8,1,1,16384,0.1213
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,1,0.1026
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,4,0.1022
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,8,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,16,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,32,0.1021
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,64,0.1009
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,128,0.1027
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,256,0.1037
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,512,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,1024,0.1055
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,2048,0.1162
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,4096,0.1192
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,8192,0.1227
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,16,1,1,16384,0.1308
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,1,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,4,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,8,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,16,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,32,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,64,0.1008
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,128,0.1018
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,256,0.1047
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,512,0.1050
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,1024,0.1069
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,2048,0.1229
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,32,1,1,4096,0.1234
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,1,0.0967
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,4,0.0945
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,8,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,16,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,32,0.0947
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,64,0.0946
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,128,0.0954
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,256,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,512,0.1028
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,1024,0.1070
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,2048,0.1355
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,64,1,1,4096,0.1397
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,1,0.0986
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,4,0.0969
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,8,0.0982
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,16,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,32,0.0968
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,64,0.0988
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,128,0.1007
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,256,0.1029
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,512,0.1117
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,1024,0.1297
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,128,1,1,2048,0.1526
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,1,0.1101
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,4,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,8,0.1083
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,16,0.1090
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,32,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,64,0.1111
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,128,0.1174
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,256,0.1271
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,512,0.1406
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,256,1,1,1024,0.1573
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,1,0.1336
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,4,0.1315
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,8,0.1334
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,16,0.1333
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,32,0.1363
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,64,0.1395
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,128,0.1522
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,256,0.1604
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,512,1,1,512,0.1770
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1024,1,1,1,0.1871
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1024,1,1,4,0.1852
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1024,1,1,8,0.1873
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1024,1,1,16,0.1878
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1024,1,1,32,0.1870
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1024,1,1,64,0.1881
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1024,1,1,128,0.2072
SGLang,0.0.0.dev1+ga4cf2ea12,NVIDIA B200,dsa_generation_module,dsa_nsa,zai-org/GLM-5,GlmMoeDsaForCausalLM,float16,fp8,fp8_block,8,1024,1,1,256,0.2214
