MMMU
11.5k college-level questions across 30 subjects requiring image + text reasoning (charts, diagrams, medical scans, music notation, …).
Best results
Frontier over time
All results
| # | Model | Score | Conditions | Eval date | Source | Flags |
|---|---|---|---|---|---|---|
| 1 | GPT 5.1 | 84.2% | 0-shot · CoT | Nov 13, 2025 | self reported | primary |
| 2 | GPT 5 (Thinking) | 84.2% | — | Aug 7, 2025 | self reported | primary |
| 3 | o3 | 82.9% | — | Apr 16, 2025 | self reported | primary |
| 4 | o4 mini | 81.6% | — | Apr 16, 2025 | self reported | primary |
| 5 | Claude Opus 4.5 | 80.7% | — | Nov 24, 2025 | self reported | primary |
| 6 | Claude Sonnet 4.5 | 77.8% | — | Sep 29, 2025 | self reported | primary |
| 7 | o1 | 77.6% | — | Apr 16, 2025 | self reported | primary |
| 8 | Qwen 3.5 122B A10B | 76.9% | — | Apr 24, 2026 | third party | primary verified |
| 9 | Llama 4 Behemoth | 76.1% | — | Apr 5, 2025 | self reported | primary |
| 10 | GPT 4.1 | 75.0% | — | Apr 14, 2025 | self reported | primary |
| 11 | Claude Sonnet 3.7 (Thinking) | 75.0% | — | Feb 24, 2025 | self reported | primary |
| 12 | Claude Sonnet 4 | 74.4% | — | May 22, 2025 | self reported | primary |
| 13 | GPT 5 | 74.4% | — | Aug 7, 2025 | self reported | primary |
| 14 | Seed 1.5 | 73.9% | — | Jan 22, 2025 | self reported | primary |
| 15 | Llama 4 Maverick | 73.4% | — | Apr 5, 2025 | self reported | primary |
| 16 | Claude Haiku 4.5 | 73.2% | — | Oct 15, 2025 | self reported | primary |
| 17 | Grok 3 | 73.2% | — | Feb 19, 2025 | self reported | primary |
| 18 | Claude Haiku 4.5 | 73.2% | — | Oct 15, 2025 | self reported | primary |
| 19 | Grok 3 | 73.2% | — | Feb 19, 2025 | self reported | primary |
| 20 | Gemini 2.5 Flash-Lite | 72.9% | — | Sep 26, 2025 | self reported | primary |
| 21 | Claude Sonnet 3.7 | 71.8% | — | Feb 24, 2025 | self reported | primary |
| 22 | Llama 4 Scout | 69.4% | — | Apr 5, 2025 | self reported | primary |
| 23 | Grok 3 mini | 69.4% | — | Feb 19, 2025 | self reported | primary |
| 24 | GPT-4o | 69.1% | — | Apr 16, 2025 | self reported | primary |
| 25 | Pixtral Large | 64.0% | CoT | Nov 18, 2024 | self reported | primary |
| 26 | Pixtral 12B | 52.0% | CoT | Oct 10, 2024 | self reported | primary |
