Skip to content

Benchmark

Design Arena Elo (models/logo)

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

Models measured
50
Scores recorded
51
Sources
1
IQ weight
1.17

Score distribution

1,176.9
Mean
1,181
Median
80.2
Std dev
1,060
10th percentile
1,268
90th percentile
1,011
Min
1,0111,365
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.17 for image-generation models, measured over 48 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Showing 51–51 of 51 results.