OpenAI says GPT-5.4 helped improve a medicinal chemistry reaction

OpenAI reported that it connected GPT-5.4 to Molecule.one's Maria, an agentic chemistry AI integrated with a high-throughput lab, and asked the system to improve an important reaction class. The best proposal focused on primary sulfonamide Chan-Lam coupling and identified TEMPO and related mild oxidants as useful additives. Across two experimental cycles, measured yields improved for 88% of boronic acids and 83% of sulfonamides tested, with mean yield rising from 16.6% to 25.2%. Human chemists repeated representative reactions at bench scale and saw higher yields for 11 of 14 substrate pairs.

Key details: Published June 17, 2026, Maria Lab ran 10,080 reactions for the OAI-M1-03 proposal, Mean yield rose from 16.6% to 25.2% under optimized conditions, Bench-scale validation improved yield for 11 of 14 substrate pairs.

Why it matters: This is a concrete lab-validated result from an AI-assisted research loop, which is much more meaningful than another benchmark-only science claim.