OpenAI introduced BioBasicsBench, a 100-task benchmark designed to evaluate the ability of AI agents to interpret raw experimental data and perform everyday tasks in wet-lab science. This benchmark assesses AI systems' capacity to extract insights and make decisions based on biological data, such as gene expression, protein structures, and cellular behavior. The tasks cover a range of common lab procedures, including data interpretation, hypothesis generation, and experiment design. BioBasbsBench is intended to help developers and researchers evaluate the performance and limitations of AI agents in wet-lab settings, facilitating the development of more accurate and reliable AI systems for biotechnology and biomedical research.