OpenCompass Agent - T-Eval - English (CompassBench 2405): leaderboard
OpenCompass CompassBench 2405 edition (file dated 2024-06-12): general tool utilization (T-Eval) sub-task of the agent 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. 39 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 88.1 | #337 |
| 2 | GPT-4 Turbo | 86.6 | #417 |
| 3 | Mixtral 8x22B Instruct | 86.2 | #738 |
| 4 | Mistral Large | 86.2 | #713 |
| 5 | Claude 3 Opus | 85.9 | #565 |
| 6 | DeepSeek V2 Chat | 85.2 | #572 |
| 7 | Yi Large | 85.1 | #566 |
| 8 | Command-R+ | 83.1 | #869 |
| 9 | Yi 1.5 34B Chat | 81.4 | #798 |
| 10 | Moonshot v1 8K | 81.2 | #622 |
| 11 | Nanbeige2-16B-Chat | 81.2 | #971 |
| 12 | GLM-4 | 80.7 | #647 |
| 13 | Qwen 1.5 32B Chat | 79.3 | #820 |
| 14 | GPT-3.5 Turbo | 78.4 | #858 |
| 15 | Yi-1.5-9B Chat | 77.6 | #1032 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-agent-t-eval-english-compassbench-2405 · How It Works · Data refreshed daily, snapshot 2026-11-05.