BlueBench - Reasoning: leaderboard
Metric: Score (%). Source: huggingface.co. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O1 | 79 |
| 2 | GPT-4o | 78 |
| 3 | GPT-4.1 | 77.5 |
| 4 | O4 Mini | 75.5 |
| 5 | Mistral Medium 3 | 75.5 |
| 6 | Mistral Large | 75 |
| 7 | Llama 3.3 70B Instruct | 73.5 |
| 8 | Pixtral-12B | 73.5 |
| 9 | GPT-4.1 Mini | 73 |
| 10 | O3 Mini | 73 |
| 11 | GPT-4.1 Nano | 69 |
| 12 | Granite 3.3 8B Instruct | 62 |
| 13 | Llama 3.2 3B Instruct | 56 |
| 14 | Llama 3.2 1B Instruct | 41 |
Interactive version: theaggregate.ai/benchmark?slug=bluebench-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-05.