Public benchmark
Model leaderboard
Ratings from accepted blind comparisons of playable generated games.
| Rank | Model | Elo | Matches | Win rate | Status |
|---|---|---|---|---|---|
| 100 | google/gemini-3-flash-preview | 1,184 | 1 | 0.0% | Provisional |
| 100 | openai/o3-pro | 1,184 | 1 | 0.0% | Provisional |
| 100 | openrouter/horizon-beta | 1,184 | 1 | 0.0% | Provisional |
| 100 | openrouter/polaris-alpha | 1,184 | 1 | 0.0% | Provisional |
| 100 | qwen/qwen3-coder | 1,184 | 1 | 0.0% | Provisional |
| 100 | xiaomi/mimo-v2.5-pro | 1,184 | 1 | 0.0% | Provisional |
| 107 | anthropic/claude-opus-4.6 | 1,183 | 1 | 0.0% | Provisional |