Benchmark
MVEB(v1) - VideoCentricQA
Vision benchmarks measure how well models understand images. Scores are percentages.
- Models measured
- 3
- Scores recorded
- 3
- Sources
- 1
- IQ weight
- No weight published
Score distribution
- 35%
- Mean
- 45.2%
- Median
- 15.4%
- Std dev
- 13.2%
- 10th percentile
- 46.6%
- 90th percentile
- 13.2%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Jina Embeddings v5 Omni SmallVideoCentricQA· MTEB Multimodal Boards data docs | 46.6% |
| 2 | Jina Embeddings v5 Omni NanoVideoCentricQA· MTEB Multimodal Boards data docs | 45.2% |
| 3 | Qwen2.5-Omni 7BVideoCentricQA· MTEB Multimodal Boards data docs | 13.2% |