OpenCompass Code - Execution - Chinese (CompassBench 2507): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 28 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Grok 4 | 32.4 |
| 2 | DeepSeek V3 (0324) | 31.1 |
| 3 | O3 (2025-04-16) (High) | 29.1 |
| 4 | Gemini 2.5 Pro | 27.7 |
| 5 | Qwen 3 235B A22B 2507 (Thinking) | 27 |
| 6 | GLM-4.5 | 27 |
| 7 | Hunyuan-T1-20250711 | 27 |
| 8 | GPT-4.1 (2025-04-14) | 26.4 |
| 9 | GLM 4.5 Air | 26.4 |
| 10 | Qwen 3 235B A22B 2507 Instruct | 26.4 |
| 11 | DeepSeek R1 0528 | 26.4 |
| 12 | MiniMax M1 80k | 26.4 |
| 13 | Kimi K2 | 25.7 |
| 14 | Llama 4 Maverick Instruct | 25.7 |
| 15 | O4 Mini (2025-04-16) (High) | 25.7 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-code-execution-chinese-compassbench-2507 · How It Works · Data refreshed daily, snapshot 2026-09-19.