Breaking

OpenAI Introduces BioBasicsBench, a 100-Task Test of AI Agents in Wet Lab Science

November 6, 2026 at 14:28 EST

Research & Papers
Healthcare

OpenAI has unveiled BioBasicsBench, a new benchmark that evaluates AI agents on 100 real-world wet lab science tasks, a test designed to measure how well such systems can carry out hands-on laboratory work rather than only answer science questions.

OpenAI · AI agents and laboratory science

BioBasicsBench tests whether AI agents can do hands-on wet lab work

OpenAI has unveiled a benchmark of 100 real-world wet lab science tasks, built to measure how well AI agents perform physical laboratory work rather than only answering science questions.

100

wet lab science tasks in the benchmark

Tasks involve experimental work with physical samples, reagents and lab equipment. The set is meant to reflect real-world laboratory activity.

Scale of the benchmark, drawn to proportion

Each square = 1 task · 100 squares = 100 tasks

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.