PetQA-Bench - Text Helpfulness: leaderboard
PetQA-Bench text-only answers rated for helpfulness on a 1-5 rubric. Higher is better. Source: arxiv.org/abs/2609.04598.
Metric: Rubric Score (1-5). Source: arxiv.org. Saturation forecast: Estimated already saturated. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 4.92 | #159 |
| 2 | Gemini 2.5 Flash | 4.91 | #255 |
| 3 | GPT-4.1 | 4.88 | #258 |
| 4 | Gemma 3 27B | 4.87 | #606 |
| 5 | GPT-4.1 Mini | 4.73 | #363 |
| 6 | Gemma 3 12B | 4.56 | #674 |
| 7 | Gemma 2 27B | 4.55 | #803 |
| 8 | Qwen 3 32B | 4.43 | #449 |
| 9 | A.X-4.0-Light | 4.21 | #776 |
| 10 | Gemma 2 9B | 4.18 | #924 |
| 11 | Qwen 3 8B | 3.75 | #678 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=petqa-bench-text-helpfulness · How It Works · Data refreshed daily, snapshot 2026-11-05.