Benchmark
MIEB(lite) - VisualSTS(eng)
Vision benchmarks measure how well models understand images. Scores are percentages.
- Models measured
- 4
- Scores recorded
- 4
- Sources
- 1
- IQ weight
- No weight published
Score distribution
- 76.5%
- Mean
- 81.5%
- Median
- 7.9%
- Std dev
- 63.2%
- 10th percentile
- 83.4%
- 90th percentile
- 63.2%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Jina Embeddings v5 Omni SmallVisualSTS(eng)· MTEB Multimodal Boards data docs | 83.4% |
| 2 | embeddinggemma-2VisualSTS(eng)· MTEB Multimodal Boards data docs | 81.5% |
| 3 | Jina Embeddings v5 Omni NanoVisualSTS(eng)· MTEB Multimodal Boards data docs | 77.9% |
| 4 | Jina Clip v1VisualSTS(eng)· MTEB Multimodal Boards data docs | 63.2% |