Event ·Official first-hand

GPT-6 Astra在OSWorld 2.0离线测试中获72.6%分

GPT-6 Astra在OSWorld 2.0离线基准测试中得分为72.6%。该测试评估模型在无互联网环境下的桌面任务执行能力。Astra能够使用计算机工具在应用程序内导航菜单、输入信息并检查屏幕结果。当任务需要时,Astra可以切换到代码模式。

Current read

OpenAI的GPT-6 Astra首次在离线桌面任务测试中突破72%大关,能自主操作应用程序并切换代码模式。

11 sources73° AI heatLast updated 2026/9/3 19:33:30

Evidence

11 sources
Related source 1The Rundown AI
View source
Related source 2Greg Brockman
View source
Related source 5Gary Marcus
View source
Related source 8歸藏(guizang.ai)
View source
Related source 9Latent Space (swyx)
View 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.