Video-MME: leaderboard
Multi-modal evaluation for video understanding covering video types, durations, and question formats.
Metric: Overall Accuracy (%). Source: video-mme.github.io. Status: saturation imminent. 51 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | video-SALMONN 2+ | 79.7 |
| 2 | Gemini 1.5 Pro | 75 |
| 3 | AdaReTaKe | 73.5 |
| 4 | Keye-VL-1.5 | 73 |
| 5 | TSPO | 73 |
| 6 | InternVL2.5 | 72.1 |
| 7 | GPT-4o | 71.9 |
| 8 | Qwen2-VL | 71.2 |
| 9 | JT-VL-Chat | 70.6 |
| 10 | LLaVA-Video | 70.6 |
Interactive version: theaggregate.ai/benchmark?slug=video-mme · How It Works · Data refreshed daily, snapshot 2026-09-05.