NextMotionQA: leaderboard
NextMotionQA (2026) tests how well vision-language models understand human motion shown as rendered 3D motion clips, with expert-verified items stratified by body part, direction, action and difficulty. This board is the overall score, the macro-average on a 0 to 100 scale of the question-answering, captioning and caption-correction task scores; higher is better.
Metric: Overall score (0-100): macro-average of the three task scores (Task 1 mean of accuracy, Jaccard and precision; Task 2 captioning mean; Task 3 correction mean) on rendered AMASS human-motion clips given to the model as RGB video; higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.6 Plus | 54.85 | #222 |
| 2 | Qwen 3.5 27B | 49.75 | #359 |
| 3 | Qwen 3.5 4B | 42.36 | #1175 |
| 4 | GPT-5.4 Mini | 42.33 | #389 |
| 5 | Qwen 3.5 9B | 42 | #803 |
| 6 | Qwen 3.5 0.8B | 20.34 | #8081 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=nextmotionqa · How It Works · Data refreshed daily, snapshot 2026-11-13.