Benchmark
MVEB(v1) - Mean (TaskType)
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
- 35.3%
- Mean
- 39.5%
- Median
- 12.4%
- Std dev
- 18.5%
- 10th percentile
- 47.9%
- 90th percentile
- 18.5%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Jina Embeddings v5 Omni SmallMean (TaskType)· MTEB Multimodal Boards data docs | 47.9% |
| 2 | Jina Embeddings v5 Omni NanoMean (TaskType)· MTEB Multimodal Boards data docs | 39.5% |
| 3 | Qwen2.5-Omni 7BMean (TaskType)· MTEB Multimodal Boards data docs | 18.5% |