Long-form video understanding from the Video-MME benchmark — questions about videos longer than 30 minutes, testing sustained attention and temporal reasoning.
This definition has not completed editorial and provenance review. Verify its metric, version, source, and evaluation configuration upstream before using it for a model-selection decision.
Tests comprehension over extended video content. Requires tracking events and details across long durations.
Requires video processing. Few evaluated models (8). Long video processing is compute-intensive.
Higher is better. Accuracy percentage (0–100%).
Requires video processing.
| # | Model | Score |
|---|---|---|
| 1 | 77.4 | |
| 2 | 77.4 | |
| 3 | 72.1 | |
| 4 | 72.1 | |
| 5 | 68.8 | |
| 6 | 68.8 | |
| 7 | 63.4 | |
| 8 | 56.9 | |
| 9 | 54.7 | |
| 10 | 47.0 |