OpenCompass LLM - Agent (CompassBench 2601): leaderboard
OpenCompass CompassBench 2601 edition (file dated 2026-02-11): agent (tool use) category score. 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: Around February 2027. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.5 (Thinking) | 76.2 | #87 (Claude Opus 4.5) |
| 2 | GLM-4.7 | 75.3 | #202 |
| 3 | Gemini 3 Pro (Preview) | 74.4 | #71 |
| 4 | Grok 4.1 | 73.1 | #239 |
| 5 | Claude Sonnet 4.5 (Thinking) | 72.6 | #148 (Claude Sonnet 4.5) |
| 6 | Kimi K2 (Thinking) | 71.4 | #252 (Kimi K2) |
| 7 | Qwen 3 Max (2026-01-23) | 70.9 | #158 |
| 8 | Tencent HY 2.0 Think | 68.9 | #206 |
| 9 | Kimi K2.5 | 68.8 | #152 |
| 10 | GPT-5.2 (2025-12-11) (High) | 68.1 | |
| 11 | DeepSeek V3.2 | 67.9 | #210 |
| 12 | Doubao-Seed-1-8-251228 | 67.7 | #153 |
| 13 | ERNIE 5.0 | 66.9 | #243 |
| 14 | MiniMax-M2.1 | 66.2 | #299 |
| 15 | O3 (2025-04-16) (High) | 60.6 | #125 (O3 (2025-04-16)) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-agent-compassbench-2601 · How It Works · Data refreshed daily, snapshot 2026-11-05.