OPENAI · 2026-11-06 19:10 UTC

Measuring how AI agents handle everyday wet-lab science

BioBasicsBench is a new benchmark, developed with biotech company Ligo, that measures how well AI agents handle routine wet-lab analysis. It has 100 tasks across five areas: protein methods, imaging and cytometry, quantification and assays, nucleic acid methods, and molecular cloning, using raw evidence such as gel images, plates, microscopy and instrument readouts. At the 'xhigh' setting, GPT-6 Astra earned an average of 66.7% of available credit versus 43.8% for GPT-5.6 Sol.

Loading the story on Tree News…