OpenCompass Code - MBPP - Chinese (CompassBench 2405): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 38 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o (2024-05-13) | 76 |
| 2 | DeepSeek V2 Chat | 69 |
| 3 | GPT-4 Turbo | 68.8 |
| 4 | Yi Large | 66.6 |
| 5 | GPT-3.5 Turbo | 63.2 |
| 6 | Llama 3 70B Instruct | 62.4 |
| 7 | Mixtral 8x22B Instruct (v0.1) | 62.4 |
| 8 | Yi-1.5-9B Chat | 57.6 |
| 9 | Command-R+ | 57 |
| 10 | Claude 3 Opus | 55.6 |
| 11 | Qwen 1.5 110B Chat | 54.8 |
| 12 | internlm2-chat-20B | 54 |
| 13 | Moonshot v1 8K | 53.8 |
| 14 | Yi 1.5 34B Chat | 53.8 |
| 15 | GLM-4 | 52.4 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-code-mbpp-chinese-compassbench-2405 · How It Works · Data refreshed daily, snapshot 2026-09-19.