Kaggle Open Benchmarks - MMLU-Pro: leaderboard
Accuracy on MMLU-Pro, a harder ten-option successor to MMLU with reasoning-focused questions across 14 subject areas. Kaggle's standardized re-run on its own harness under the open-benchmarks organization, not the original authors' leaderboard; Kaggle's fraction is shown as a percentage (0-100, higher is better). Ten answer options, so guessing scores about 10.
Metric: Accuracy (%). Source: www.kaggle.com. Saturation forecast: Estimated already saturated. 62 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon | 93.19 | #9 |
| 2 | Claude Opus 5 | 91.68 | #16 |
| 3 | Gemini 3.1 Pro (Preview) | 91.29 | #64 |
| 4 | Gemini 3.8 Flash | 90.56 | #30 |
| 5 | Gemini 3.7 Flash | 90.31 | #24 |
| 6 | GPT-6 Astra | 89.8 | #6 |
| 7 | Gemini 3.6 Flash | 89.79 | #57 |
| 8 | Gemini 3.5 Flash | 89.56 | #66 |
| 9 | Grok 4.5 | 89.45 | #54 |
| 10 | GPT-6.1 Sol | 89.44 | #7 |
| 11 | Gemini 3 Flash (Preview) | 89.42 | #90 |
| 12 | Claude Opus 4.8 | 88.66 | #39 |
| 13 | GPT-6 Sol | 88.06 | #19 |
| 14 | Claude Opus 4.1 | 87.89 | #146 |
| 15 | GPT-5 | 87.11 | #105 |
No result here: #3 Claude Opus 5.5, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=kaggle-open-benchmarks-mmlu-pro · How It Works · Data refreshed daily, snapshot 2026-11-05.