MuseBench - Multi-Select (Overall): leaderboard
MuseBench evaluates intent-level audiovisual-arts understanding in multimodal models over 4,016 questions from video essays in four art categories (cinematic, static visual, stage performing, game arts), with single-select and multi-select framings. This board reports Overall exact-match accuracy (%) on multi-select questions; higher is better.
Metric: Overall exact-match accuracy (%) on multi-select questions; higher is better. Source: arxiv.org. Saturation forecast: Around 2031. 28 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 28.91 | #102 |
| 2 | GPT-5.4 | 25.5 | #123 |
| 3 | Qwen 3.5 Plus | 23.21 | #232 |
| 4 | Qwen 3.5 397B A17B | 22.71 | #246 |
| 5 | InternVL3-8B | 20.3 | #1688 |
| 6 | Qwen2.5-Omni-7B | 18.18 | #1789 |
| 7 | Gemini 3.1 Pro (Preview) | 14.88 | #94 |
| 8 | InternVL3-78B | 13.53 | #738 |
| 9 | Gemma 4 E4B | 9.06 | #2214 |
| 10 | GLM-4.5V | 8.61 | #744 |
| 11 | Grok 4.1 | 8 | #476 |
| 12 | Kimi K2.5 | 2.07 | #256 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=musebench-multi-select-overall · How It Works · Data refreshed daily, snapshot 2026-11-13.