GeneBench-Pro

Researchers released GeneBench-Pro, a 129-problem benchmark testing AI agents on multi-stage genomics and biomedicine analyses. GPT-5.6 Sol Pro (Extended) achieved the top score of 31.5%, highlighting the difficulty of realistic scientific workflows for current AI systems.

GeneBench-Pro is a 129-problem benchmark for AI agents performing realistic multi-stage scientific analyses in genomics, quantitative biology, and translational biomedicine. Problems require agents to inspect messy datasets, choose and revise analysis workflows, and return decision-relevant quantitative answers under ambiguity. Category: Biology. Imported rows: 29. Top imported result: GPT-5.6 Sol Pro Extended , rank 1, 31.5%.