MuseBench - Game Arts (Single-Select): leaderboard
Metric: Game-arts chance-adjusted accuracy (%): raw single-select accuracy minus the per-item chance baseline, rescaled so random guessing scores 0 and a correct answer 1; a negative value is worse than chance; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 28 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 Plus | 38.8 |
| 2 | Gemini 3.1 Pro (Preview) | 38.72 |
| 3 | Qwen 3.5 397B A17B | 35.79 |
| 4 | Claude Opus 4.6 | 34.07 |
| 5 | GPT-5.4 | 32 |
| 6 | InternVL3-78B | 31.59 |
| 7 | Gemma 4 E4B | 10.3 |
| 8 | Qwen2.5-Omni-7B | 8.31 |
| 9 | InternVL3-8B | 6.66 |
| 10 | Grok 4.1 | 3.2 |
| 11 | Kimi K2.5 | 0.35 |
| 12 | GLM-4.5V | -4.34 |
Interactive version: theaggregate.ai/benchmark?slug=musebench-game-arts-single-select · How It Works · Data refreshed daily, snapshot 2026-09-29.