OpenCompass Reasoning - Common - Chinese (CompassBench 2507): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Doubao-Seed-1.6-thinking-250715 | 88.4 |
| 2 | Qwen 3 235B A22B 2507 Instruct | 83.7 |
| 3 | GPT-5 | 81.4 |
| 4 | Claude Sonnet 4 (Thinking) | 80.2 |
| 5 | ERNIE-X1-Turbo-32K | 80.2 |
| 6 | Hunyuan-T1-20250711 | 80.2 |
| 7 | DeepSeek V3 (0324) | 79.1 |
| 8 | GLM-4.5 | 79.1 |
| 9 | Grok 4 | 75.6 |
| 10 | Kimi K2 | 75.6 |
| 11 | GPT-4.1 (2025-04-14) | 73.3 |
| 12 | GLM 4.5 Air | 69.8 |
| 13 | MiniMax M1 80k | 69.8 |
| 14 | MiniMax-Text-01 | 58.1 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-reasoning-common-chinese-compassbench-2507 · How It Works · Data refreshed daily, snapshot 2026-09-19.