Event ·Multi-source

Alibaba 发布 Co-Design Bench：最强模型也只能完整解决 26.4% 的任务

Alibaba 推出编程智能体基准 Co-Design Bench，测试集全部来自真实用户请求中曾让智能体失败的任务。表现最好的 GPT-6 Astra 完整解决率只有 26.4%，意味着约四分之三的任务仍需人工收尾。而同样的模型在旧基准上的成绩超过 99%，两套数字之间的落差暴露了现有评测与实际使用之间的差距。

Current read

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。

4 sources57° AI heatLast updated 2026/11/8 09:50:59

Evidence

4 sources

Primary sourceDecoder

2026/11/8 09:50:59View source

Related source 1rohanpaul_ai

2026/11/6 19:54:08View source

Related source 2Yuchen Jin

2026/11/6 14:40:21View source

Related source 3Artificial Analysis

2026/11/6 17:09:35View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story