Skip to content

Benchmark

AHELM: Multilinguality - Mean win rate

Multilingual benchmarks compare model performance across languages. Scores are percentages. Higher is better.

Models measured
12
Scores recorded
12
Sources
1
IQ weight
0.68

Score distribution

48.1%
Mean
56.3%
Median
23%
Std dev
25%
10th percentile
72.9%
90th percentile
6.3%
Min
6.3%85.4%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.68 for chat models, measured over 9 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard