OpenAI has introduced FrontierScience, a new benchmark designed to measure high-level scientific reasoning in AI models across physics, chemistry, and biology. The system features two distinct tracks: the Olympiad set, which uses complex short-answer problems created by international medalists, and the Research set, which consists of PhD-level sub-tasks. To evaluate these open-ended research probl...去小宇宙查看完整单集简介
前往小宇宙评论区与主播互动