Skip to content

Benchmark

HealthAgentBench

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.

Models measured
10
Scores recorded
12
Sources
1
IQ weight
1.18

Score distribution

31.2%
Mean
32%
Median
11.6%
Std dev
17%
10th percentile
45%
90th percentile
16%
Min
16%55%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.18 for chat models, measured over 10 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 5clinical-result-129· Clinical Benchmarks data docs55%
2GPT-5.6 Solclinical-result-130· Clinical Benchmarks data docs45%
3GPT 5.5clinical-result-131· Clinical Benchmarks data docs42%
4Claude Opus 4.8clinical-result-132· Clinical Benchmarks data docs36%
5GPT 5.5clinical-result-133· Clinical Benchmarks data docs35%
6Claude Opus 4.8clinical-result-134· Clinical Benchmarks data docs32%
7GPT 5.4clinical-result-321· Clinical Benchmarks data docs28%
8Claude Opus 4.7clinical-result-322· Clinical Benchmarks data docs27%
9GPT-5.3 Codexclinical-result-1822· Clinical Benchmarks data docs22%
10Claude Opus 4.6clinical-result-323· Clinical Benchmarks data docs19%
11Claude Sonnet 4.6clinical-result-324· Clinical Benchmarks data docs17%
12GPT-5.4 miniclinical-result-325· Clinical Benchmarks data docs16%

Data credits: Clinical Benchmarks (clinicalbenchmarks.ai), CC BY 4.0; fractions converted to percentages; source documents retain their own terms (licence)