🚨 AI News | TestingCatalog@testingcatalog
54AI 编辑部评分,满分 100
2026-08-08 01:26· 20分钟前
AI 导读

OpenAI 内部报告称,即将推出的模型 Astra 在智能体编码和网络安全方面取得显著进展,公司据此判定其可能具备关键网络能力。OpenAI 随即对更高能力模型实施更严格的安全控制,包括隔离测试环境、限制网络和工具访问、加强模型权重保护与加密,并暂停不符合新要求的 Astra 相关内部活动。

OPENAI 🔥: A new report published by the company states that Astra, "one of their upcoming models," "indicates significant advancements in agentic coding and cybersecurity."

These results, in addition to expert assessments, have led us to conclude last night that we cannot rule out critical cyber capabilities under our Preparedness Framework.

We are implementing stricter security controls for higher-capability models and associated activities, including isolated testing environments, restricted network and tool access, enhanced model weight protections and encryption, additional monitoring and detection capabilities, and sandboxed execution.

We are pausing internal activities involving Astra that do not yet meet these strengthened security control requirements.

Astra not soon? 👀

Tibor Blaho"Our latest internal evaluations of Astra, one of our upcoming models, over the past few days indicate significant advancements in agentic coding and cybersecur...

来源:🚨 AI News | TestingCatalog · x.com

🚨 AI News | TestingCatalog · @testingcatalog · X·2026-08-08 01:26·20分钟前
AI 导读

OpenAI 内部报告称,即将推出的模型 Astra 在智能体编码和网络安全方面取得显著进展,公司据此判定其可能具备关键网络能力。OpenAI 随即对更高能力模型实施更严格的安全控制,包括隔离测试环境、限制网络和工具访问、加强模型权重保护与加密,并暂停不符合新要求的 Astra 相关内部活动。

OPENAI 🔥: A new report published by the company states that Astra, "one of their upcoming models," "indicates significant advancements in agentic coding and cybersecurity."

These results, in addition to expert assessments, have led us to conclude last night that we cannot rule out critical cyber capabilities under our Preparedness Framework.

We are implementing stricter security controls for higher-capability models and associated activities, including isolated testing environments, restricted network and tool access, enhanced model weight protections and encryption, additional monitoring and detection capabilities, and sandboxed execution.

We are pausing internal activities involving Astra that do not yet meet these strengthened security control requirements.

Astra not soon? 👀

Tibor Blaho"Our latest internal evaluations of Astra, one of our upcoming models, over the past few days indicate significant advancements in agentic coding and cybersecur...

来源:🚨 AI News | TestingCatalog· x.com