SVHalluc - Cross-Modal Temporal Binding: leaderboard
Speech-vision hallucination (SVHalluc, Zhang et al., CVPR 2026): 339 three-way questions on which visual action happens while the speech is heard, for audio-visual LLMs on cooking videos with narration. Option-balanced; accuracy in percent, higher is better.
Metric: Accuracy (%; 339 option-balanced video questions, temporal binding of speech and visual events (three options)). Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 69.25 | #262 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=svhalluc-cross-modal-temporal-binding · How It Works · Data refreshed daily, snapshot 2026-11-12.