MMLU Pro: leaderboard
Academic multiple-choice benchmark covering 14 subjects including STEM, humanities, and social sciences.
Metric: Accuracy (%). Source: vals.ai. Saturation forecast: Estimated already saturated. 138 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5.1 (Max) | 92.38 | #9 (Claude Fable 5.1) |
| 2 | Claude Opus 5 (Max) | 91.59 | #12 (Claude Opus 5) |
| 3 | Claude Fable 5 (Max) | 91.5 | #16 (Claude Fable 5) |
| 4 | Gemini 3.1 Pro (Preview) (High) | 90.99 | #58 (Gemini 3.1 Pro (Preview)) |
| 5 | Gemini 3.8 Flash (High) | 90.22 | #21 (Gemini 3.8 Flash) |
| 6 | Gemini 3.7 Flash (High) | 90.12 | #23 (Gemini 3.7 Flash) |
| 7 | Gemini 3 Pro (Preview) | 90.1 | #66 |
| 8 | Claude Opus 4.7 (Max) | 89.87 | #49 (Claude Opus 4.7) |
| 9 | Claude Opus 4.8 (Max) | 89.58 | #37 (Claude Opus 4.8) |
| 10 | Gemini 3.5 Flash (High) | 89.52 | #60 (Gemini 3.5 Flash) |
| 11 | Grok 4.6 (High) | 89.4 | #26 (Grok 4.6) |
| 12 | Qwen 3.7 Max | 89.31 | #75 |
| 13 | Gemini 3.6 Flash (High) | 89.28 | #51 (Gemini 3.6 Flash) |
| 14 | Grok 4.5 (High) | 89.22 | #48 (Grok 4.5) |
| 15 | Claude Opus 4.6 (Adaptive Reasoning, Max Effort) | 89.11 | #63 (Claude Opus 4.6) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mmlu-pro · How It Works · Data refreshed daily, snapshot 2026-10-20.