Video-MME what it measures, how it is scored and who leads it
Multimodal · Video-MME team · introduced May 31, 2024 · Accuracy
About this benchmark
- Task
- Answer questions about videos from 11 seconds to an hour long, with or without subtitles and audio.
- Dataset
- Videos across 6 visual domains and 30 subfields, with human-written multiple-choice questions.
- Method
- Multiple choice, reported by video length and with or without subtitles.
- Metric
- Accuracy
- Organization
- Video-MME team
- Introduced
- May 31, 2024
- Official leaderboard
- video-mme.github.io/
Versions
Newest first. New versions are added, never rewritten.
| Version | Date | |
|---|---|---|
| 1Released with the paper. | May 31, 20242 years ago | Released with the paper. |
Sources: each benchmark's paper (arXiv) and its official site or leaderboard. Scores are the boards captured here every day, exactly as published.