VEBench - Footage Selection: leaderboard
Metric: Footage Selection accuracy (%) on Operation Simulation: four-option choice of the candidate clip that best continues a reference video (A/B-roll pairing), candidates concatenated into one sequence with on-screen text markers, subtitles provided; higher is better. Source: arxiv.org. Saturation forecast: Around October 2027. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 44.44 |
| 2 | GPT-4o | 41.89 |
| 3 | Qwen 3 VL 32B Instruct | 39.72 |
| 4 | Qwen 3 VL 8B Instruct | 36.14 |
| 5 | Qwen 3 VL 4B Instruct | 31.8 |
| 6 | InternVL3-14B | 31.16 |
| 7 | InternVL3-8B | 30.78 |
| 8 | Qwen 2.5 VL 7B Instruct | 30.01 |
| 9 | Qwen 3 VL 30B A3B Instruct | 29.37 |
Interactive version: theaggregate.ai/benchmark?slug=vebench-footage-selection · How It Works · Data refreshed daily, snapshot 2026-10-07.