Anthropic 让 Claude 模型自主完成蛋白质设计，命中率超行业水平
原标题：Anthropic says any lab can now let a language model agent run the whole protein design stack
AI 摘要
Anthropic 发布两项实验，让 Claude 模型（Mythos Preview 和 Opus 4.8）自主运行蛋白质设计流程，在早期药物发现中设计微型结合蛋白。结果显示，Claude 在 15 个目标中成功 14 个，实验室测试命中率达 26.8%，优于行业通常的 10%-15%。模型使用开源工具如 RFdiffusion 和 SolubleMPNN，并通过 Adaptyv Bio 和 Twist Bioscience 验证，部分结果优于竞赛获胜设计。独立评审尚未进行。
正文节选
Anthropic says any lab can now let a language model agent run the whole protein design stack In two experiments, Anthropic put its Claude models to work on early-stage drug discovery. The company says its protein design results beat the usual industry hit rates. An independent review of the results is still pending. Anthropic has published two experiments in which its Claude models took on tasks from the early phase of drug development. The first focused on designing so-called minibinders, small