BeQu - Experiment 1 - Entailment Precision: leaderboard
Metric: Entailment Precision (%). Source: arxiv.org. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 Nano (Medium) | 88 |
| 2 | Gemini 3.1 Pro (Preview) (Medium) | 79.2 |
| 3 | GPT-5.4 (Medium) | 76.6 |
| 4 | Claude Sonnet 4.6 (Medium) | 64.6 |
| 5 | Llama 4 Scout Instruct | 63.2 |
| 6 | GPT-5 Mini (Medium) | 62.6 |
| 7 | Claude Haiku 4.5 | 61.4 |
| 8 | MiniMax-M2.5 | 60.8 |
| 9 | Claude Opus 4.6 (Medium) | 59.6 |
| 10 | Kimi K2.5 | 54.4 |
| 11 | Mistral Large 3 | 53.4 |
| 12 | Grok 4.1 Fast | 53.2 |
| 13 | Qwen 3.5 27B | 49 |
| 14 | Gemini 3 Flash (Preview) (Medium) | 47.8 |
| 15 | DeepSeek V3.2 | 46 |
Interactive version: theaggregate.ai/benchmark?slug=bequ-experiment-1-entailment-precision · How It Works · Data refreshed daily, snapshot 2026-09-19.