Kaggle Open Benchmarks - AIME 2025: leaderboard
Accuracy on the 30 problems of the 2025 American Invitational Mathematics Examination (AIME I and II), integer answers from 0 to 999 graded by exact match. 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). With 30 problems, each question is worth about 3.3 points.
Metric: Accuracy (%). Source: www.kaggle.com. Saturation forecast: Estimated already saturated. 66 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 100 | #33 |
| 2 | Claude Opus 4.8 | 100 | #39 |
| 3 | Claude Opus 5 | 100 | #16 |
| 4 | GPT-6 Astra | 100 | #6 |
| 5 | Claude Opus 5.5 | 100 | #3 |
| 6 | GPT-6.1 Sol | 100 | #7 |
| 7 | Gemini 4 Argon | 100 | #9 |
| 8 | Grok 4.5 | 100 | #54 |
| 9 | GPT-6 Sol | 100 | #19 |
| 10 | Grok 4.7 | 100 | #26 |
| 11 | GPT-5.6 Sol | 96.67 | #22 |
| 12 | Gemini 3.5 Flash | 96.67 | #66 |
| 13 | DeepSeek V4.1 Flash | 96.67 | #37 |
| 14 | GPT-6 Luna | 96.67 | #61 |
| 15 | Claude Sonnet 5.5 | 96.67 | #14 |
No result here: #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=kaggle-open-benchmarks-aime-2025 · How It Works · Data refreshed daily, snapshot 2026-11-05.