Skip to content

Benchmark

Vals BioMysteryBench

Science benchmarks measure how well models answer science questions. Scores are percentages.

Models measured
33
Scores recorded
33
Sources
1
IQ weight
2.27

Score distribution

70.2%
Mean
70%
Median
8.2%
Std dev
61.5%
10th percentile
79.6%
90th percentile
42.6%
Min
42.6%83.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.27 for chat models, measured over 33 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard