Event ·Multi-source

OpenAI称Astra为最危险模型

OpenAI将即将发布的Astra模型评为首个具备'关键'网络能力的系统。该公司计划通过监控思维链来控制该模型。然而，这种监控已被证明是模型真实决策的不可靠反映。据报告，Astra的新架构将更多思维过程推向不可读区域。随着能力提升，安全网可能正在变弱。

Current read

OpenAI的Astra模型被标记为最危险系统，新架构让监控变得更困难，安全与能力平衡面临挑战。

11 sources73° AI heatLast updated 2026/9/2 14:20:28

Evidence

11 sources

Primary sourceDecoder

2026/9/2 14:20:28View source

Related source 1OpenAI

2026/9/1 20:30:17View source

Related source 2techcrunch

2026/9/2 20:19:14View source

Related source 3Gary Marcus

2026/9/3 19:43:20View source

Related source 4VnExpress Số hóa

2026/9/2 14:58:03View source

Related source 5Geek

2026/9/3 06:48:29View source

Related source 6The Rundown AI

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

Related source 7Greg Brockman

2026/9/3 19:44:36View source

Related source 8Sam Altman

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

Related source 9lmarena.ai

2026/9/3 20:26:06View source

Related source 10Latent.Space

2026/9/3 21:12:41View 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