Skip to content

Benchmark

MMEB VisDoc - ViDoRe_arxivqa

Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages.

Models measured
1
Scores recorded
1
Sources
1
IQ weight
No weight published

Score distribution

87%
Mean
87%
Median
0%
Std dev
87%
10th percentile
87%
90th percentile
87%
Min

Leaderboard

#ModelScore
1Qwen3-VL Embedding 8BViDoRe_arxivqa· MMEB (TIGER-Lab) data87%

Data credits: MMEB Leaderboard (TIGER-Lab), Apache-2.0 (licence)