OpenCompass Multimodal - MMMU: leaderboard
OpenCompass multimodal score on MMMU multidisciplinary visual-question tasks.
Metric: Score (%). Source: rank.opencompass.org.cn. Saturation forecast: Around January 2027. 105 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 81.8 | #98 |
| 2 | GPT-5 Mini (2025-08-07) | 78.7 | #172 |
| 3 | Gemini 2.5 Pro | 74.7 | #152 |
| 4 | GPT-4.1 (2025-04-14) | 74 | #250 |
| 5 | GPT-4o ChatGPT | 72.9 | |
| 6 | GPT-5 Nano | 72.6 | #425 |
| 7 | InternVL3-78B | 72.2 | #352 |
| 8 | GPT-4.5 | 72.1 | #274 |
| 9 | Doubao-1.5-Pro | 71.4 | #287 |
| 10 | Claude 3.7 Sonnet | 71 | #249 |
| 11 | Gemini 2.0 Flash | 69.9 | #342 |
| 12 | InternVL3-38B | 69.7 | #405 |
| 13 | Grok 2 (1212) | 67.1 | #496 |
| 14 | InternVL3-14B | 64.8 | #505 |
| 15 | Gemma 3 27B | 64.8 | #607 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-multimodal-mmmu · How It Works · Data refreshed daily, snapshot 2026-10-20.