OpenCompass Reasoning - Common (CompassBench 2601): leaderboard
OpenCompass CompassBench 2601 edition (file dated 2026-02-11): common reasoning sub-task of the reasoning category. 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 | Gemini 3 Pro (Preview) | 73.6 | #71 |
| 2 | GPT-5.2 (2025-12-11) (High) | 69.6 | |
| 3 | Claude Opus 4.5 (Thinking) | 68.2 | #87 (Claude Opus 4.5) |
| 4 | Kimi K2.5 | 64.5 | #152 |
| 5 | DeepSeek V3.2 | 61.2 | #210 |
| 6 | Claude Sonnet 4.5 (Thinking) | 60.8 | #148 (Claude Sonnet 4.5) |
| 7 | GLM-4.7 | 60.8 | #202 |
| 8 | Grok 4.1 | 60.8 | #239 |
| 9 | Qwen 3 Max (2026-01-23) | 60.8 | #158 |
| 10 | O3 (2025-04-16) (High) | 59.9 | #125 (O3 (2025-04-16)) |
| 11 | GPT-OSS-120B (High) | 59.7 | #346 (GPT-OSS-120B) |
| 12 | Tencent HY 2.0 Think | 59.6 | #206 |
| 13 | Ring-1T | 56.9 | #326 |
| 14 | Qwen 3 235B A22B 2507 (Thinking) | 56.1 | #263 |
| 15 | ERNIE 5.0 | 55.5 | #243 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-reasoning-common-compassbench-2601 · How It Works · Data refreshed daily, snapshot 2026-11-05.