OpenAI introduces a 100-task benchmark for wet-lab AI

OpenAI has introduced BioBasicsBench, a 100-task benchmark for AI agents interpreting routine wet-lab data.

GPT-6 Astra earned 66.7% of available credit at the highest reported reasoning setting, compared with 43.8% for GPT-5.6 Sol.

The tasks test data interpretation, error detection and whether evidence supports a conclusion.