Skip to content

Benchmark

Design Arena Elo (models/svg)

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

Models measured
100
Scores recorded
112
Sources
1
IQ weight
1.97

Score distribution

1,139.3
Mean
1,140
Median
115.3
Std dev
977
10th percentile
1,288
90th percentile
856
Min
8561,413
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.97 for chat models, measured over 98 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Showing 101–112 of 112 results.