Event ·Multi-source

GPT6-Astra评测：编程能力与Claude Opus 5相当

GPT6-Astra在Coding Agent能力上与Claude Opus 5、Fable 5相当，但落后于Fable 5.1。综合智能指数部分维度甚至落后于GPT5.6。Token效率高，Computer Use等特定任务表现优秀。ARC-AGI-3在OpenAI特殊适配Harness上达到99%，标准跑分仅63%。模型不显示推理过程时能完成更复杂任务，但带来安全风险。

Current read

OpenAI的GPT6-Astra评测曝光，编程能力与Claude Opus 5相当，但价格涨2.5倍，因效率提升实际成本反而降一半。

11 sources73° AI heatLast updated 2026/9/4 08:57:07

Evidence

11 sources

Primary source向阳乔木

2026/9/4 08:57:07View source

Related source 1Simon Willison’s Weblog

2026/9/3 20:18:41View source

Related source 2歸藏(guizang.ai)

2026/9/4 00:59:38View source

Related source 3orange.ai

2026/9/3 23:36:56View source

Related source 4宝玉

2026/9/5 02:05:42View source

Related source 5lmarena.ai

2026/9/5 17:40:41View source

Related source 6Aravind Srinivas

2026/9/3 21:13:22View source

Related source 7shao__meng

2026/9/3 23:29:45View source

Related source 8Gary Marcus

2026/9/4 01:34:52View source

Related source 9Decoder

2026/9/4 11:07:36View source

Related source 10Perplexity

2026/9/4 22:43:47View 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