Skip to content
InfoResearchIndustryLLM-specific

Can AI help scientists identify promising drug candidates?

Published
Record updated
View JSON

Summary

MedChemBench is a benchmark that evaluates AI models on reasoning tasks drawn from medicinal chemistry research, covering molecular structures, scientific reasoning, compound ranking and selection, and property predictions. It uses 99 problems drawn from 117 papers and patents, with 761 distinct molecular structures. At the xhigh reasoning setting, GPT-6.2 Astra earned 84.5% of available credit, compared with 58.2% for GPT-5.6 Sol.