OpenVLM MMMU - Music: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 284 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 46.7 |
| 2 | Claude 3.7 Sonnet | 46.7 |
| 3 | GPT-4.1 (2025-04-14) | 46.7 |
| 4 | Gemini 2.5 Pro | 43.3 |
| 5 | InternVL3-38B | 43.3 |
| 6 | Gemini 1.5 Pro (002) | 43.3 |
| 7 | Grok 2 (1212) | 43.3 |
| 8 | GPT-4o ChatGPT | 43.3 |
| 9 | GPT-4.5 | 43.3 |
| 10 | InternVL2-8B | 40 |
| 11 | Qwen 2 VL 72B | 40 |
| 12 | GPT-5 Nano | 40 |
| 13 | InternVL2.5-78B | 40 |
| 14 | Claude 3 Haiku | 40 |
| 15 | Llama 3.2 90B Vision Instruct | 40 |
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmmu-music · How It Works · Data refreshed daily, snapshot 2026-09-05.