# RewardHarness endpoint-list schema v1: one OpenAI-compatible URL per line.
# Blank lines and lines beginning with # are ignored.
# Each URL serves Qwen2.5-VL-7B-Instruct on a different GPU/host.
#
# Examples (replace with your own hosts/ports):
http://localhost:8000/v1
http://localhost:8001/v1
http://localhost:8002/v1
http://localhost:8003/v1
# http://my-gpu-host-01:8000/v1
# http://my-gpu-host-01:8001/v1
