OpenAI 发布 GPT-6 Cyber：可攻破系统也能提出修复方案
原标题：OpenAIs GPT-6 Cyber can both break into systems and propose how to fix them
AI 摘要
OpenAI 发布 GPT-6 Cyber，一款专门用于发现软件安全漏洞、构建可用漏洞利用并建议补丁的 AI 模型。在真实测试中，该模型在 Firefox 中发现 81 个此前未知的严重漏洞（Mozilla 已修复），并在 Linux 内核中复现 1,724 个漏洞、为其中 948 个开发出可用利用，数量超过前代五倍。模型目前仅通过 API 和 Codex 向部分 Daybreak Red 客户开放，OpenAI 同时强调滥用风险上升并列出多项防护措施。
正文节选
OpenAIs GPT-6 Cyber can both break into systems and propose how to fix them Key Points - OpenAI has unveiled GPT-6 Cyber, an AI model specifically trained to find security vulnerabilities, develop working exploits, and suggest patches. - In targeted tests, GPT-6 Cyber found 81 previously unknown critical vulnerabilities in Firefox, which Mozilla has already patched. - In the Linux kernel, the model reproduced 1,724 vulnerabilities and developed working exploits for 948 of them, more than five ti