Anthropic’s BioMysteryBench tests Claude on real bioinformatics data, not standardized exams. Claude matched human experts on solvable problems and outperformed them on 23 questions designed to stump specialists.
Anthropic’s new benchmark shows Claude matches and sometimes outperforms human experts in bioinformatics
Source: Complete Ai Training
Read Full Story →
