BenchCompass - Context-Grounded Reasoning (Open Book): leaderboard
Metric: Accuracy (%). Source: arxiv.org. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 89.6 |
| 2 | Qwen 3.5 397B A17B | 84.4 |
| 3 | Qwen 3.5 122B A10B | 84.4 |
| 4 | GPT-5.5 | 82.3 |
| 5 | Gemma 4 31B (IT) | 82.3 |
| 6 | Claude Opus 4.7 | 82.3 |
| 7 | Gemma 4 26B A4B (IT) | 82.3 |
| 8 | DeepSeek V4 Pro | 81.3 |
| 9 | MiniMax-M2.5 | 75 |
| 10 | GLM-5.1 | 74 |
| 11 | Qwen 3 32B | 69.8 |
| 12 | Llama 4 Maverick | 68.8 |
| 13 | Mistral Large 3 | 64.6 |
| 14 | Qwen 3 8B | 58.3 |
Interactive version: theaggregate.ai/benchmark?slug=benchcompass-context-grounded-reasoning-open-book · How It Works · Data refreshed daily, snapshot 2026-09-20.