ODU-Bench (Audio-Visual) - User Profile: leaderboard

Metric: Profile accuracy (%; mean per-field accuracy of predicted user-in-frame, gender and age group (gender and age group without video), demand scenes only; 1,347 audio-visual scenes (1,061 with a demand, 286 without; 1,070 synthetic and 277 human-recorded), user turns as video with audio, earlier assistant replies as text). Source: arxiv.org. Saturation forecast: Around December 2026. 12 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)95.7
2Qwen 3.5 Omni Plus93.5
3Qwen3 Omni 30B A3B Instruct92.8
4Seed 2.0 Lite92.5
5Gemini 3.7 Flash88.3
6Gemini 3.5 Flash Lite79.3
7Qwen2.5-Omni-7B78.2
8MiniCPM-o-4.568.1

Interactive version: theaggregate.ai/benchmark?slug=odu-bench-audio-visual-user-profile · How It Works · Data refreshed daily, snapshot 2026-09-26.