Skip to content

Benchmark

PGG-Bench (Lech Mazur)

Scores are percentages.

Models measured
17
Scores recorded
18
Sources
1
IQ weight
1.17

Score distribution

10%
Mean
10%
Median
2%
Std dev
8.3%
10th percentile
11.9%
90th percentile
5.1%
Min
5.1%14.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.17 for chat models, measured over 17 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard