Skip to content

Benchmark

MedHELM: MTSamples Procedures - Jury Score

Scores are percentages. Higher is better.

Models measured
13
Scores recorded
13
Sources
1
IQ weight
1.74

Score distribution

4%
Mean
4%
Median
0.452%
Std dev
3.6%
10th percentile
4.5%
90th percentile
2.8%
Min
2.8%4.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.74 for chat models, measured over 13 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

Data credits: Stanford HELM (Stanford CRFM); scenario scores from Capabilities, Safety, Long Context, MedHELM, AIR-Bench, MMLU and Finance