OpenCompass LLM - Agent (CompassBench 2409): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 30 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.1 405B Instruct FP8 | 87.7 |
| 2 | Llama 3.1 70B Instruct | 86.6 |
| 3 | Yi Large | 86.2 |
| 4 | GPT-4o Mini (2024-07-18) | 86 |
| 5 | Gemini 1.5 Pro | 85.7 |
| 6 | Step 2 16K | 85.4 |
| 7 | Gemma 2 27B (IT) | 85.2 |
| 8 | Mistral Large 2 (Jul) | 85.2 |
| 9 | GPT-4o (2024-08-06) | 85.1 |
| 10 | DeepSeek V2.5 | 85 |
| 11 | GPT-4o (2024-05-13) | 84.3 |
| 12 | Mistral-Small-Instruct-2409 | 84 |
| 13 | Qwen 2.5 72B Instruct | 83.8 |
| 14 | Gemma 2 9B (IT) | 82.3 |
| 15 | Claude 3.5 Sonnet (20240620) | 82.1 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-agent-compassbench-2409 · How It Works · Data refreshed daily, snapshot 2026-09-19.