OpenCompass LLM - Knowledge (CompassBench 2601): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 26 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 95.5 |
| 2 | Kimi K2.5 | 92.8 |
| 3 | O3 (2025-04-16) (High) | 92.7 |
| 4 | DeepSeek V3.2 | 92.5 |
| 5 | Claude Opus 4.5 (Thinking) | 92.2 |
| 6 | GPT-5.2 (High) | 92 |
| 7 | Tencent HY 2.0 Think | 91.8 |
| 8 | Claude Sonnet 4.5 (Thinking) | 91.2 |
| 9 | GLM-4.7 | 91.1 |
| 10 | Qwen 3 Max (2026-01-23) | 90.9 |
| 11 | Qwen 3 235B A22B 2507 (Thinking) | 90.3 |
| 12 | Ling-1T | 89.8 |
| 13 | ERNIE 5.0 | 89.5 |
| 14 | Ring-1T | 88.4 |
| 15 | Kimi K2 (Thinking) | 88.2 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-knowledge-compassbench-2601 · How It Works · Data refreshed daily, snapshot 2026-09-19.