BlueBench - Legal: leaderboard
Metric: Score (%). Source: huggingface.co. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Mistral Medium 3 | 67.17 |
| 2 | GPT-4o | 60.87 |
| 3 | GPT-4.1 | 60.47 |
| 4 | Llama 3.3 70B Instruct | 59.54 |
| 5 | GPT-4.1 Mini | 59.32 |
| 6 | GPT-4.1 Nano | 56.91 |
| 7 | O4 Mini | 56.4 |
| 8 | Granite 3.3 8B Instruct | 55.94 |
| 9 | O1 | 51.22 |
| 10 | Llama 3.2 3B Instruct | 50.56 |
| 11 | O3 Mini | 37.87 |
| 12 | Llama 3.2 1B Instruct | 37.26 |
| 13 | Mistral Large | 34.71 |
| 14 | Pixtral-12B | 32.9 |
Interactive version: theaggregate.ai/benchmark?slug=bluebench-legal · How It Works · Data refreshed daily, snapshot 2026-09-05.