OpenCompass Code - MBPP - English (CompassBench 2405): leaderboard
OpenCompass CompassBench 2405 edition (file dated 2024-06-12): code generation (sanitized MBPP) sub-task of the code category, English questions. 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: Estimated already saturated. 38 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 87.2 | #337 |
| 2 | Yi Large | 79 | #566 |
| 3 | GPT-4 Turbo | 77.4 | #417 |
| 4 | GPT-3.5 Turbo | 77 | #858 |
| 5 | Llama 3 70B Instruct | 75.9 | #640 |
| 6 | Mixtral 8x22B Instruct | 75.5 | #738 |
| 7 | Claude 3 Opus | 75.5 | #565 |
| 8 | GLM-4 | 74.3 | #647 |
| 9 | Yi-1.5-9B Chat | 73.5 | #1032 |
| 10 | DeepSeek V2 Chat | 73.2 | #572 |
| 11 | Yi 1.5 34B Chat | 72.8 | #798 |
| 12 | Yi-1.5-6B-Chat | 69.3 | #1241 |
| 13 | Moonshot v1 8K | 67.3 | #622 |
| 14 | Command-R+ | 66.2 | #869 |
| 15 | internlm2-chat-20B | 65.8 | #988 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-code-mbpp-english-compassbench-2405 · How It Works · Data refreshed daily, snapshot 2026-11-05.