FitAQA - Judgement: leaderboard
Metric: Form-error F1 (%; error-positive class, judging whether an exercise aspect is executed correctly; over FitAQA's fitness action-quality videos (30 bodyweight exercises, 38-error taxonomy); judgement task). Source: arxiv.org. Saturation forecast: Around December 2026. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 72.1 |
| 2 | GPT-5.5 | 57.8 |
| 3 | GPT-5.4 | 50.5 |
| 4 | Qwen 3.5 27B (Non-reasoning) | 47.1 |
| 5 | Qwen 3.5 9B (Non-reasoning) | 45.7 |
| 6 | Qwen 3.5 27B (Thinking) | 42.6 |
| 7 | Qwen 3.5 9B (Thinking) | 39.9 |
| 8 | Qwen 3.5 35B A3B (Thinking) | 39 |
| 9 | Qwen 3.5 35B A3B (Non-reasoning) | 33.4 |
| 10 | Gemma 4 31B | 27.8 |
| 11 | Qwen 3 VL 8B | 26.8 |
Interactive version: theaggregate.ai/benchmark?slug=fitaqa-judgement · How It Works · Data refreshed daily, snapshot 2026-09-26.