AA MMMU-Pro: leaderboard
Artificial Analysis independent evaluation of MMMU-Pro: multimodal understanding across 30 academic disciplines, eliminating shortcut strategies.
Metric: Accuracy (%). Source: artificialanalysis.ai. Status: saturation imminent. 258 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6 (Max) | 86.88 |
| 2 | GPT-6 (High) | 86.42 |
| 3 | GPT-6 (xHigh) | 86.24 |
| 4 | Gemini 3.8 Flash (High) | 85.61 |
| 5 | Gemini 3.7 Flash (High) | 85.49 |
| 6 | Median Expert | 85.4 |
| 7 | GPT-6 (Medium) | 85.09 |
| 8 | Gemini 3.7 Flash (Low) | 84.86 |
| 9 | Claude Opus 5 (Adaptive Reasoning, Max Effort) | 84.74 |
| 10 | Gemini 3.7 Flash (Medium) | 84.74 |
| 11 | GPT-6 (Low) | 84.62 |
| 12 | Gemini 3.8 Flash (Low) | 84.51 |
| 13 | Gemini 3.5 Flash (High) | 84.28 |
| 14 | Gemini 3.8 Flash (Medium) | 84.22 |
| 15 | Claude Opus 5 (Adaptive Reasoning, Xhigh Effort) | 83.99 |
Interactive version: theaggregate.ai/benchmark?slug=aa-mmmu-pro · How It Works · Data refreshed daily, snapshot 2026-09-05.