Skip to content

Benchmark

MIEB(lite) - DocumentUnderstanding

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

57.6%
Mean
74.2%
Median
24.9%
Std dev
14.8%
10th percentile
74.3%
90th percentile
14.8%
Min
14.8%74.3%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Jina Embeddings v5 Omni SmallDocumentUnderstanding· MTEB Multimodal Boards data docs74.3%
2Jina Embeddings v5 Omni NanoDocumentUnderstanding· MTEB Multimodal Boards data docs74.2%
3embeddinggemma-2DocumentUnderstanding· MTEB Multimodal Boards data docs67.1%
4Jina Clip v1DocumentUnderstanding· MTEB Multimodal Boards data docs14.8%