MEGA-Bench Task - Video Eval Visual Pref — leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.5 Sonnet (20241022) | 87.5 |
| 2 | Gemma 3 27B (IT) | 81.2 |
| 3 | Gemini 2.5 Pro | 75 |
| 4 | Claude 3.5 Sonnet (20240620) | 75 |
| 5 | Gemma 3 12B (IT) | 68.8 |
| 6 | Qwen 2 VL 72B | 68.8 |
| 7 | GPT-4o | 62.5 |
| 8 | Gemma 3 4B (IT) | 56.2 |
| 9 | InternVL2-8B | 56.2 |
| 10 | Qwen 2 VL 7B | 56.2 |
| 11 | Pixtral-12B | 56.2 |
| 12 | InternVL2.5-78B | 56.2 |
| 13 | InternVL2.5-2B | 56.2 |
| 14 | Gemini 2.0 Flash (Preview) | 56.2 |
| 15 | InternVL3-8B | 50 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-video-eval-visual-pref · How the rankings work · Data refreshed daily, snapshot 2026-07-22.