Skip to content

Benchmark

MMEB VisDoc - VisRAG_MP-DocVQA

Vision benchmarks measure how well models understand images. Scores are percentages.

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

Score distribution

89.3%
Mean
89.3%
Median
0%
Std dev
89.3%
10th percentile
89.3%
90th percentile
89.3%
Min

Leaderboard

#ModelScore
1Qwen3-VL Embedding 8BVisRAG_MP-DocVQA· MMEB (TIGER-Lab) data89.3%

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