Benchmark
MVEB(v1)
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
- 31.5%
- Mean
- 36.2%
- Median
- 13.8%
- Std dev
- 12.8%
- 10th percentile
- 45.5%
- 90th percentile
- 12.8%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Jina Embeddings v5 Omni Small· MTEB Multimodal Boards data docs | 45.5% |
| 2 | Jina Embeddings v5 Omni Nano· MTEB Multimodal Boards data docs | 36.2% |
| 3 | Qwen2.5-Omni 7B· MTEB Multimodal Boards data docs | 12.8% |