BlueBench - Summarization: leaderboard
BlueBench summarization task covering congressional bills and long social posts with reference summaries.
Metric: Score (%). Source: huggingface.co. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.3 70B Instruct | 19.47 | #534 |
| 2 | GPT-4.1 | 18.25 | #248 |
| 3 | Mistral Large | 18.24 | #710 |
| 4 | Llama 3.2 3B Instruct | 17.99 | #1352 |
| 5 | GPT-4.1 Mini | 17.92 | #355 |
| 6 | Llama 3.2 1B Instruct | 17.62 | #1539 |
| 7 | Mistral Medium 3 | 17.43 | #500 |
| 8 | Pixtral-12B | 17.28 | #808 |
| 9 | Granite 3.3 8B Instruct | 17.15 | #825 |
| 10 | GPT-4o | 17.06 | #343 |
| 11 | GPT-4.1 Nano | 16.94 | #731 |
| 12 | O1 | 16.77 | #183 |
| 13 | O3 Mini | 16.66 | #272 |
| 14 | O4 Mini | 16.3 | #180 |
Interactive version: theaggregate.ai/benchmark?slug=bluebench-summarization · How It Works · Data refreshed daily, snapshot 2026-10-20.