OpenAI推出GPT-5.5-Cyber并上线网络信任访问框架

Odaily星球日报讯 OpenAI 正式推出 GPT-5.5-Cyber 模型及面向网络安全防御者的“网络信任访问”(Trusted Access for Cyber,简称 TAC)框架。同时,GPT-5.5-Cyber 已向负责关键基础设施的防御者开放限量预览,以支持专业化网络安全工作流。
TAC 是一个基于身份和信任的框架,旨在确保增强的 AI 能力由获验证的防御者掌握。通过该框架验证的防御者在执行漏洞识别、分诊、恶意软件分析、二进制逆向工程及补丁验证等任务时,将减少触发模型拒绝响应的情况。自 2026 年 6 月 1 日起,访问该能力的个人成员将被要求启用高级账户安全保护。
OpenAI 目前正与 Cisco、CrowdStrike、Palo Alto Networks 等安全厂商合作,通过 GPT-5.5 加速安全生态系统的防御循环,提升漏洞研究、修补、监测及供应链安全的效率。
Disclaimer: OKX Orbit content is provided for informational purposes only. Learn more
Replies
Related Flash News
Kevin Walsh will be sworn in at the White House on Friday, with Trump presiding over the ceremony
Tempo integrates with the DeFi lending marketplace Morpho
Minnesota signed bill allowing banks to offer crypto custody services
Anza recommended Solana mainnet validators to upgrade Agave 4, and the Turbine retransmission delay dropped to about 0.8ms
Yunshen plans to raise about 2.503 billion yuan in IPO to invest in embodied algorithms and model research and development
The S&P 500 index and Dow Jones Industrial Average quickly turned lower at the beginning of the session
Citi: Quantum computing breakthroughs are accelerating, and Bitcoin faces excess quantum risks
The net BTC purchases of listed companies rebounded strongly to $2.03 billion in a single week, a surge of more than 44 times from last week
Strategy's total floating profit on Bitcoin narrowed to $1.35 billion, and BitMNR's floating loss on Ethereum expanded to $7.279 billion
Vitalik Buterin: Ethereum develops a new paradigm and formal verification


