Skip to content

Benchmark

MVEB(text, video, v1) - VideoPairClassification

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

75.3%
Mean
85.6%
Median
13.5%
Std dev
53.1%
10th percentile
86.8%
90th percentile
53.1%
Min
53.1%86.8%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Qwen3-VL Embedding 8BVideoPairClassification· MTEB Multimodal Boards data docs86.8%
2Jina Embeddings v5 Omni SmallVideoPairClassification· MTEB Multimodal Boards data docs85.6%
3Jina Embeddings v5 Omni NanoVideoPairClassification· MTEB Multimodal Boards data docs75.8%
4Qwen2.5-Omni 7BVideoPairClassification· MTEB Multimodal Boards data docs53.1%