Tag
1 article
LifeSciBench is a comprehensive benchmark developed by OpenAI that evaluates AI models on real-life science research tasks, focusing on reasoning and decision-making rather than simple recall. It uses expert-authored rubrics to assess how well AI systems can handle complex scientific workflows.
