MEGA-Bench Task - Video Segments Reordering: leaderboard

Metric: Task Score (%). Source: huggingface.co. 44 models tracked.

Top models

#ModelScore
1InternVL3-38B14.3
2Gemini 2.5 Pro (Preview 03-25)14.3
3GPT-4o7.1
4Claude 3.5 Sonnet (20240620)7.1
5Gemma 3 27B (IT)0
6Gemma 3 12B (IT)0
7Gemma 3 4B (IT)0
8InternVL3-78B0
9InternVL3-8B0
10InternVL2-8B0
11Qwen 2 VL 72B0
12Qwen 2 VL 7B0
13Claude 3.5 Sonnet (20241022)0
14Qwen 2 VL 2B0
15Aquila-VL-2B0

Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-video-segments-reordering · How It Works · Data refreshed daily, snapshot 2026-09-05.