OpenAI and biotechnology firm Ligo have launched BioBasicsBench. The new benchmark features 100 real experimental tasks to measure how effectively AI agents interpret raw wet-lab data and catch scientific errors.
FireBrief
FireBrief