Skip to content

Benchmark

Vals CUA-bench

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

Models measured
8
Scores recorded
8
Sources
1
IQ weight
0.45

Score distribution

10%
Mean
9%
Median
4.8%
Std dev
4.8%
10th percentile
14%
90th percentile
4.2%
Min
4.2%19.2%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.45 for chat models, measured over 8 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard