Benchmark
MVEB(text, video, v1) - Any2AnyRetrieval
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
- 39.7%
- Mean
- 52.6%
- Median
- 25.4%
- Std dev
- 0.39%
- 10th percentile
- 69.2%
- 90th percentile
- 0.39%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Qwen3-VL Embedding 8BAny2AnyRetrieval· MTEB Multimodal Boards data docs | 69.2% |
| 2 | Jina Embeddings v5 Omni SmallAny2AnyRetrieval· MTEB Multimodal Boards data docs | 52.6% |
| 3 | Jina Embeddings v5 Omni NanoAny2AnyRetrieval· MTEB Multimodal Boards data docs | 36.6% |
| 4 | Qwen2.5-Omni 7BAny2AnyRetrieval· MTEB Multimodal Boards data docs | 0.39% |