Important new eval!OpenAI (@OpenAI):We’re releasing a new eval to measure expert-level scientific reasoning: FrontierScience.This benchmark measures PhD-level scientific reasoning across physics, chemistry, and biology.It contains hard, expert-written questions (both olympiad-style problems and longer ...