MEDit-Bench: leaderboard
Metric: mIoU (%; temporal IoU between the union of the predicted cuts and the union of the reference cuts, averaged over the 540 video, message and reference-edit triplets (60 long-form videos, three editing messages each, three professional edits per message); zero-shot: the model maps the video and the message to a sequence of cuts). Source: arxiv.org. Saturation forecast: Around 2031. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 25.3 |
| 2 | Gemini 3 Pro | 23.5 |
| 3 | Qwen 3 VL 32B | 22.7 |
| 4 | GPT-5 Mini | 18.6 |
| 5 | Gemma 4 E4B (Non-reasoning) | 16.9 |
| 6 | Gemini 2.5 Pro | 15.2 |
| 7 | InternVL3.5-8B | 8 |
Interactive version: theaggregate.ai/benchmark?slug=medit-bench · How It Works · Data refreshed daily, snapshot 2026-09-29.