BlueBench - Bias: leaderboard
Metric: Score (%). Source: huggingface.co. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4.1 | 97.98 |
| 2 | O3 Mini | 97.98 |
| 3 | Mistral Large | 96.97 |
| 4 | Llama 3.3 70B Instruct | 95.96 |
| 5 | GPT-4.1 Mini | 94.95 |
| 6 | O4 Mini | 94.95 |
| 7 | O1 | 94.95 |
| 8 | GPT-4o | 93.94 |
| 9 | Mistral Medium 3 | 93.94 |
| 10 | GPT-4.1 Nano | 81.82 |
| 11 | Granite 3.3 8B Instruct | 76.77 |
| 12 | Pixtral-12B | 72.73 |
| 13 | Llama 3.2 3B Instruct | 60.61 |
| 14 | Llama 3.2 1B Instruct | 48.48 |
Interactive version: theaggregate.ai/benchmark?slug=bluebench-bias · How It Works · Data refreshed daily, snapshot 2026-09-05.