OpenCompass LLM - Math (CompassBench 2601): leaderboard
OpenCompass CompassBench 2601 edition (file dated 2026-02-11): math category score. CompassBench is OpenCompass's periodic evaluation of chat models on a private question set; each edition is a separate board. Score 0-100, higher is better.
Metric: Score (%). Source: rank.opencompass.org.cn. Saturation forecast: Around February 2027. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 Max (2026-01-23) | 73.2 | #158 |
| 2 | Kimi K2.5 | 72.9 | #152 |
| 3 | Gemini 3 Pro (Preview) | 71.8 | #71 |
| 4 | GPT-5.2 (2025-12-11) (High) | 71.4 | |
| 5 | Kimi K2 (Thinking) | 70.3 | #252 (Kimi K2) |
| 6 | GLM-4.7 | 69.7 | #202 |
| 7 | Tencent HY 2.0 Think | 69.4 | #206 |
| 8 | Claude Opus 4.5 (Thinking) | 69.2 | #87 (Claude Opus 4.5) |
| 9 | DeepSeek V3.2 | 68.7 | #210 |
| 10 | Ring-1T | 68.3 | #326 |
| 11 | GPT-OSS-120B (High) | 68.1 | #346 (GPT-OSS-120B) |
| 12 | Claude Sonnet 4.5 (Thinking) | 68 | #148 (Claude Sonnet 4.5) |
| 13 | Qwen 3 235B A22B 2507 (Thinking) | 67.1 | #263 |
| 14 | Qwen 3 Next 80B A3B (Thinking) | 65.9 | #312 |
| 15 | Doubao-Seed-1-8-251228 | 65.8 | #153 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-math-compassbench-2601 · How It Works · Data refreshed daily, snapshot 2026-11-05.