Event ·Multi-source

OpenAI发布GPT-6 Astra群星模型

OpenAI今天凌晨公布了全新架构的GPT-6 Astra群星模型，在科学研究和抽象推理方面大幅提升。该模型在Terminal-Bench Science 0.1基准测试中从22.4%提升至64.6%，大幅领先Fable的52.6%；在FrontierMath Tier 4(v2)达到97.6%。Astra在动作与步数效率(ARC-AGI-3)上比人类中位数减少51.7%的动作步骤，但在编程能力DeepSWE v1.1上仅得74.1%，与Claude Opus 5相当。

Current read

OpenAI的GPT-6 Astra在科研推理上大幅领先，但编程能力提升有限，与竞品差距不大。

11 sources93° AI heatLast updated 2026/9/3 23:36:56

Evidence

11 sources

Primary sourceorange.ai

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

Related source 1Greg Brockman

2026/9/3 20:20:14View source

Related source 2elvis

2026/9/3 19:55:42View source

Related source 3The Rundown AI

2026/9/3 18:07:36View source

Related source 4OpenAI

2026/9/3 19:32:16View source

Related source 5Sam Altman

2026/9/3 19:49:10View source

Related source 6Stanford AI Lab

2026/9/3 21:56:12View source

Related source 7歸藏(guizang.ai)

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

Related source 8Gary Marcus

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

Related source 9向阳乔木

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

Related source 10Decoder

2026/9/4 17:23:35View 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