VideoZeroBench - Temporal Grounding: leaderboard
Metric: Mean multi-segment temporal IoU (%) between the evidence intervals the model predicts from the video and question and the annotated intervals (0 for a missing or invalid interval or a question without temporal evidence), on VideoZeroBench's 500 manually annotated open-ended questions (exact short answers) over 138 long videos in 13 domains; Gemini models get the raw video through the API, the others 1 fps frames up to the printed cap, temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around May 2027. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro | 32 |
| 2 | Gemini 2.5 Pro | 31.4 |
| 3 | Gemini 2.5 Flash | 27.9 |
| 4 | Seed 2.0 Pro | 25.3 |
| 5 | GPT-5.2 | 21.9 |
| 6 | Qwen 3 VL 235B A22B | 19.6 |
| 7 | Qwen 3.5 397B A17B | 3.1 |
| 8 | Qwen 2.5 VL 7B | 1.8 |
Interactive version: theaggregate.ai/benchmark?slug=videozerobench-temporal-grounding · How It Works · Data refreshed daily, snapshot 2026-10-07.