ModelsBuildersMulti-source18:11Chinese contentAlibaba 发布 Co-Design Bench:最强模型也只能完整解决 26.4% 的任务Alibaba 搞了个新基准,专挑真实翻车任务来测,最强的 GPT-6 Astra 也只过四分之一,跟旧基准 99% 的成绩反差很大。#Co-Design Bench#Alibaba#GPT-6 Astra#编程助手3 个信源在谈Event官方账号Decoder@Jonathan Kemper4 sourcesSourceRead laterReadWorth followingUsefulFollow Co-Design Bench