BlueBench - Product Help: leaderboard
BlueBench product-help task using consumer-finance complaint data to classify and reason about support issues.
Metric: Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.3 70B Instruct | 86.64 | #534 |
| 2 | O3 Mini | 86.08 | #272 |
| 3 | GPT-4o | 85.95 | #343 |
| 4 | GPT-4.1 Mini | 84.95 | #355 |
| 5 | O4 Mini | 84.27 | #180 |
| 6 | O1 | 83 | #183 |
| 7 | GPT-4.1 | 82.33 | #248 |
| 8 | Mistral Medium 3 | 79.65 | #500 |
| 9 | Llama 3.2 3B Instruct | 75.58 | #1352 |
| 10 | Granite 3.3 8B Instruct | 73.88 | #825 |
| 11 | GPT-4.1 Nano | 68.52 | #731 |
| 12 | Mistral Large | 67.02 | #710 |
| 13 | Pixtral-12B | 59.87 | #808 |
| 14 | Llama 3.2 1B Instruct | 45.62 | #1539 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=bluebench-product-help · How It Works · Data refreshed daily, snapshot 2026-10-20.