TraceAV-Bench - Audio-to-Visual Deception: leaderboard
Metric: Accuracy (%) on the 229 Audio-to-Visual Deception (A2V) questions of the multimodal hallucination dimension, four-option multiple choice over long audio-visual videos (10 to 140 minutes), exact match of the selected option set, each model at its official inference setting, with the full audio and video input; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.0 Flash | 81.66 |
| 2 | Gemini 3.1 Pro (Preview) | 79.91 |
| 3 | Gemini 3 Flash | 75.55 |
| 4 | Gemini 2.5 Pro | 74.24 |
| 5 | Gemma 4 E4B | 69.43 |
| 6 | Gemini 2.5 Flash | 66.81 |
| 7 | Qwen2.5-Omni-7B | 55.46 |
Interactive version: theaggregate.ai/benchmark?slug=traceav-bench-audio-to-visual-deception · How It Works · Data refreshed daily, snapshot 2026-10-07.