Skip to content

Benchmark

Design Arena Elo (models/website)

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

Models measured
148
Scores recorded
161
Sources
1
IQ weight
1.85

Score distribution

1,153.6
Mean
1,172
Median
128.5
Std dev
983
10th percentile
1,300
90th percentile
737
Min
7371,374
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.85 for chat models, measured over 148 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Showing 151–161 of 161 results.