TraceAV-Bench - Audio-to-Visual Deception: leaderboard

The Audio-to-Visual Deception sub-task of TraceAV-Bench (2026), in its multimodal hallucination dimension: questions that claim a sight from what is heard in long audio-visual videos. The score is accuracy in percent with full audio and video input; higher is better.

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 February 2027. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 2.0 Flash81.66#723
2Gemini 3.1 Pro (Preview)79.91#99
3Gemini 3 Flash75.55#175
4Gemini 2.5 Pro74.24#276
5MiniCPM-o-4.572.05#1252
6Qwen 3 Omni 30B A3B69.87#847
7Gemma 4 E4B69.43#2184
8Ming-Flash-Omni-2.067.25#1266
9Gemini 2.5 Flash66.81#463
10Qwen2.5-Omni-7B55.46#1884
11OmniVinci-9B44.1
12Baichuan-Omni-1.541.48#3285
13video-SALMONN 239.3#2687
14HumanOmni-7B37.99#3308
15VideoLLaMA2.1-7B-AV29.69#4204

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=traceav-bench-audio-to-visual-deception · How It Works · Data refreshed daily, snapshot 2026-11-21.