OpenCompass Code - TACO (CompassBench 2405): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 39 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Moonshot v1 8K | 19.6 |
| 2 | GPT-4o (2024-05-13) | 13.4 |
| 3 | Claude 3 Opus | 13 |
| 4 | GPT-4 Turbo | 11.4 |
| 5 | GPT-3.5 Turbo | 9.8 |
| 6 | DeepSeek V2 Chat | 9 |
| 7 | Yi Large | 8.2 |
| 8 | Llama 3 70B Instruct | 8 |
| 9 | GLM-4 | 6.2 |
| 10 | Mistral Large | 6 |
| 11 | Mixtral 8x22B Instruct (v0.1) | 5.5 |
| 12 | DBRX | 5.4 |
| 13 | Command-R+ | 5.1 |
| 14 | Qwen 1.5 110B Chat | 5 |
| 15 | Qwen 1.5 32B Chat | 4.8 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-code-taco-compassbench-2405 · How It Works · Data refreshed daily, snapshot 2026-09-19.