ReLE - Reasoning and Mathematics: leaderboard
Metric: Accuracy (%). Source: nonelinear.com. 174 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 5 | 92 |
| 2 | Claude Opus 4.8 (Thinking) | 89.9 |
| 3 | Qwen 3.8 Flash | 87.7 |
| 4 | Kimi K3 | 86.3 |
| 5 | Seed 2.0 Lite | 85.8 |
| 6 | Qwen 3.5 122B A10B | 85.5 |
| 7 | Gemini 3.1 Pro (Preview) | 85.1 |
| 8 | GLM-5.3 Flash | 84.8 |
| 9 | GPT-5.2 (High) | 84.8 |
| 10 | Qwen 3.7 Max | 84.7 |
| 11 | GPT-5.1 (High) | 84.7 |
| 12 | Gemini 3.5 Flash | 84.5 |
| 13 | Qwen 3.7 Plus | 84.5 |
| 14 | GLM-5.3 | 84.5 |
| 15 | Gemini 3.8 Flash | 83.9 |
Interactive version: theaggregate.ai/benchmark?slug=rele-reasoning-and-mathematics · How It Works · Data refreshed daily, snapshot 2026-09-19.