Skip to content

Benchmark

Arena ELO Vision - Captioning

People compare model answers and choose which they prefer. Higher scores are better.

Models measured
30
Scores recorded
33
Sources
1
IQ weight
1.77

Score distribution

1,215.8
Mean
1,212.6
Median
63.5
Std dev
1,144.5
10th percentile
1,287.4
90th percentile
1,058.5
Min
1,058.51,343.1
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.77 for chat models, measured over 30 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Data credits: LMArena, CC BY 4.0 (licence)