VideoOdyssey-AV - Temporal Grounding: 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 locating when an event happens; higher is better.
Metric: Multiple-choice accuracy (%) on the 61 temporal grounding questions (locate the timestamp of specific audio-visual events) 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
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 Omni Plus | 42.6 | #254 |
| 2 | Gemini 3.1 Pro (Preview) | 39.3 | #99 |
| 3 | Gemini 2.5 Pro | 37.7 | #276 |
| 4 | Qwen 3 Omni 30B A3B | 37.7 | #847 |
| 5 | Gemini 3 Flash | 32.8 | #175 |
| 6 | Ola-7B | 32.8 | #1557 |
| 7 | VideoLLaMA2-7B | 31.2 | #2385 |
| 8 | Unified-IO-2 XXL | 31.2 | |
| 9 | OneLLM | 24.6 | #2372 |
| 10 | Unified-IO-2 XL | 21.3 | |
| 11 | Unified-IO-2 L | 19.7 | |
| 12 | VITA-1.5-7B | 13.1 | #1853 |
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-temporal-grounding · How It Works · Data refreshed daily, snapshot 2026-11-21.