OpenAI 发布药物化学基准 MedChemBench
OpenAI has now published MedChemBench, its medicinal chemistry benchmark.
OpenAI 发布药物化学基准 MedChemBench,题目取自 117 篇论文和专利,涵盖 761 个评分分子结构,并隐去研究人员后来报告的结果。每个提示词停在真实决策节点,提供生物靶点、已有化合物及部分测定结果。表现最好的 GPT-6.2 Astra 总体得分约 84%,读取化学信息接近满分,但在关键决策上约有一半判断正确。
推荐理由
AI 评分 53
该基准把药物化学任务截断在真实决策节点,可用于区分模型读取化学信息与作出关键判断的能力。
主题
正文
OpenAI has now published MedChemBench, its medicinal chemistry benchmark.
Its problems come from 117 papers and patents, span 761 graded molecular structures, and withhold the results researchers reported later.
Each prompt stops at a real decision point, supplying the biological target, the compounds made so far and selected assay results.
the best model GPT-6.2 Astra gets about 84% overall, is near perfect at reading the chemistry, and is right about half the time on the decision that actually matters.
AI 热点