MotionHalluc - Attributional Hallucination: leaderboard
Choosing which of two corrective instructions names the right joint and direction of the motion difference (600 questions). MotionHalluc (Guo et al., 2026) tests whether multimodal models compare a query and a reference fitness-exercise video (553 pairs from Fit3D) without hallucinating motion differences; open models decode greedily and API models are averaged over three runs. Accuracy in percent; higher is better.
Metric: Accuracy (%; choose the one of two corrective instructions that names the right joint and direction; 600 questions). Source: arxiv.org. Saturation forecast: Around February 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 83.5 | #159 |
| 2 | Qwen 3.5 Plus | 81.72 | #227 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=motionhalluc-attributional-hallucination · How It Works · Data refreshed daily, snapshot 2026-11-12.