← bench
GameDevBench
Play and compare LLM-generated games
Local 30B
OmniCoder
Sonnet
Qwopus 27B
Qwen2.5-Coder
Haiku
Opus
GPT-5.4-pro
Gemini Flash
Gemini Pro
muLLM Router
WITH RICE ⚡
10 games
20 games
50 games
All games
▶ Run
Filter:
All
Local Won
Cloud Won
Local Failed
Interesting
Both Failed
Games
🎮
Select a game from the list
Run benchmarks to generate games, then compare results side-by-side
⚙ Running benchmarks…
hide (runs continue in background)