HypoEvolve wraps a genetic algorithm around multi-agent LLMs for drug repurposing.

Thirteen authors ran a generational genetic algorithm over specialized agents that separately handle mechanistic argument, assumption reconsideration, and evidence and testability assessment (arXiv 2609.15938). Evaluated against DepMap and Open Targets across 34 cancer types, it reached 0.171 DepMap selectivity against 0.115 for the strongest baseline, beating six baselines on both measures. Gains over single-pass generation held on held-out cancer types too, which is the part that matters: the evolutionary loop carried the improvement, not agent specialization alone.