Skip to content

Benchmark

Vals SRE Bench

Scores are percentages.

Models measured
34
Scores recorded
34
Sources
1
IQ weight
4.34

Score distribution

12.3%
Mean
4.2%
Median
15.8%
Std dev
0%
10th percentile
33.6%
90th percentile
0%
Min
0%56.9%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 4.34 for chat models, measured over 34 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard