Skip to content

Benchmark

MMMU - Science

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

Models measured
2
Scores recorded
2
Sources
1
IQ weight
No weight published

Score distribution

37.6%
Mean
38.5%
Median
0.9%
Std dev
36.7%
10th percentile
38.5%
90th percentile
36.7%
Min
36.7%38.5%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1Pixtral 12BScience, 2409· EvalAI data38.5%
2LLaVA 1.6 34BScience· EvalAI data36.7%