Kaggle Open Benchmarks - MATH-500: leaderboard
Accuracy on MATH-500, a 500-problem subset of the MATH competition-mathematics dataset spanning seven subjects and five difficulty levels, graded by final-answer equivalence. 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).
Metric: Accuracy (%). Source: www.kaggle.com. Saturation forecast: Estimated already saturated. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 5.5 | 99.41 | #14 |
| 2 | Gemini 3.8 Flash | 99.11 | #30 |
| 3 | Grok 4.6 | 99.08 | #29 |
| 4 | GPT-6.1 Sol | 99.01 | #7 |
| 5 | Gemini 3.7 Flash | 99 | #24 |
| 6 | Gemini 4 Argon | 99 | #9 |
| 7 | Claude Haiku 5.5 | 98.71 | #43 |
| 8 | GPT-6 Sol | 98.64 | #19 |
| 9 | Claude Opus 5 | 97.61 | #16 |
| 10 | GPT-5.6 Sol | 96.65 | #22 |
| 11 | Gemini 3.5 Flash | 96.39 | #66 |
| 12 | GPT-5.6 Terra | 95.9 | #40 |
| 13 | Gemini 3.6 Flash | 95.86 | #57 |
| 14 | Gemini 3 Flash (Preview) | 94.11 | #90 |
| 15 | GPT-5.6 Luna | 93.69 | #79 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=kaggle-open-benchmarks-math-500 · How It Works · Data refreshed daily, snapshot 2026-11-05.