OpenAI and Ligo introduce BioBasicsBench to test how AI agents read raw wet-lab data

OpenAI on November 6 introduced BioBasicsBench, a benchmark built with Ligo, an enzyme-design biotech company. Its 100 everyday wet-lab tasks test whether AI agents can interpret raw experimental data, spot errors and recognize when evidence cannot support a conclusion. The tasks span protein methods, imaging and cytometry, quantification and assays, nucleic acid methods and molecular cloning, using gel images, plate photos, microscopy, sequencing reads and bench videos from scientists' own experiments, each scored by an expert-written rubric. Of 7 models tested, GPT-6 Astra earned 66.7% of the available credit at the xhigh setting (OpenAI's own figure). OpenAI also lists failures, such as letting an assumption override the controls and overcounting colonies on one plate by about 12%.

Written by dalao.ai: the title, summary and why it matters are our own words. The original reports are the record; their publishers keep authorship.