OpenCompass Knowledge - Engineering (CompassBench 2601): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 26 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 95.8 |
| 2 | DeepSeek V3.2 | 95.8 |
| 3 | Kimi K2.5 | 94.8 |
| 4 | GPT-5.2 (High) | 94.5 |
| 5 | Claude Opus 4.5 (Thinking) | 94.5 |
| 6 | O3 (2025-04-16) (High) | 93.5 |
| 7 | Claude Sonnet 4.5 (Thinking) | 93 |
| 8 | GLM-4.7 | 92.2 |
| 9 | Qwen 3 Max (2026-01-23) | 92.2 |
| 10 | ERNIE 5.0 | 91.8 |
| 11 | Qwen 3 235B A22B 2507 (Thinking) | 91.2 |
| 12 | Ling-1T | 91 |
| 13 | Tencent HY 2.0 Think | 90.8 |
| 14 | Kimi K2 (Thinking) | 89.8 |
| 15 | Grok 4.1 | 89.2 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-knowledge-engineering-compassbench-2601 · How It Works · Data refreshed daily, snapshot 2026-09-19.