VideoOdyssey-AV - Acoustic Attribute Recognition: leaderboard

VideoOdyssey (2026) tests ultra-long video understanding with human-written multiple-choice questions on videos averaging 109 minutes, each annotated with how long a viewer must watch continuously to answer. This board is accuracy in percent on the audio-visual subset questions about recognizing sound attributes; higher is better.

Metric: Multiple-choice accuracy (%) on the 65 acoustic attribute recognition questions (identify the acoustic attributes of specific acoustic events, based on both visual and audio cues) of VideoOdyssey-AV (human-written questions on 100 videos of 60 to 255 minutes that need both visual and audio cues; 1,062 in all), omni-modal models given video and audio, 64 frames (VITA-1.5-7B 16); higher is better. Source: arxiv.org. Saturation forecast: Around July 2027. 12 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3.5 Omni Plus52.3#254
2Gemini 3 Flash46.2#175
3Gemini 2.5 Pro44.6#276
4Gemini 3.1 Pro (Preview)41.5#99
5Ola-7B36.9#1557
6VideoLLaMA2-7B32.3#2385
7Unified-IO-2 XXL32.3
8Qwen 3 Omni 30B A3B29.2#847
9Unified-IO-2 XL29.2
10VITA-1.5-7B29.2#1853
11OneLLM21.5#2372
12Unified-IO-2 L21.5

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

Interactive version: theaggregate.ai/benchmark?slug=videoodyssey-av-acoustic-attribute-recognition · How It Works · Data refreshed daily, snapshot 2026-11-21.