内容来源：https://www.theverge.com/ai-artificial-intelligence/928342/openai-daybreak-security-ai

内容总结：

OpenAI发布“破晓”网络安全计划：AI自动寻找并修补漏洞，对标对手“神话”系统

美国人工智能公司OpenAI近日正式推出名为“Daybreak”（破晓）的网络安全倡议，旨在利用AI技术在黑客发动攻击前自动检测并修复系统漏洞。该计划的核心是其今年3月发布的Codex Security AI智能体，能够根据企业的代码库建立威胁模型，分析潜在攻击路径，验证高危漏洞，并实现自动化优先处理。

值得注意的是，Daybreak并非单一模型驱动，而是整合了OpenAI最强模型能力，包括上周刚刚上线的GPT-5.5-Cyber及具备可信访问权限的网络安全专用版本。OpenAI还透露，正与行业及政府合作伙伴联手，准备“逐步部署能力更强的网络模型”。

此次发布距其主要竞争对手Anthropic推出“Claude Mythos”（神话）系统仅一个多月。Anthropic当时声称该安全模型因风险过高不宜公开发布，仅通过其“Glasswing”（玻璃翼）计划秘密向部分用户开放，但仍有未授权方设法获取了访问权限。

在此之前，OpenAI在这一领域的产品布局相对滞后。如今，随着Daybreak的推出，双方在AI网络安全攻防战中的竞争正式进入白热化阶段。

中文翻译：

OpenAI正式推出Daybreak，这是一项专注于在攻击者发现漏洞之前进行检测与修复的人工智能计划。Daybreak利用今年3月发布的Codex安全智能体，基于组织的代码构建威胁模型，聚焦可能的攻击路径，验证潜在漏洞，并自动检测风险较高的漏洞。

OpenAI刚刚发布了其应对Claude Mythos的解决方案
OpenAI的Daybreak整合了GPT-5.5-Cyber与Codex安全系统。
OpenAI的Daybreak整合了GPT-5.5-Cyber与Codex安全系统。

就在一个多月前，竞争对手Anthropic宣布推出Claude Mythos，这是一款专注于安全的人工智能模型，该公司声称该模型过于危险而无法公开发布，仅作为其名为"Project Glasswing"的计划的一部分进行私下分享。尽管如此，仍有至少少数未经授权的团体获得了访问权限。

然而，OpenAI此前一直缺乏类似的安全产品。与Glasswing一样，Daybreak并非仅基于单一AI模型构建——OpenAI表示"Daybreak整合了最强大的OpenAI模型、Codex以及我们的安全合作伙伴。"

Daybreak还涉及专门的网络模型，包括上周开始推出的具备可信访问权限的GPT-5.5-Cyber及GPT-5.5-Cyber。OpenAI还表示，在准备"部署能力越来越强的网络模型"的同时，正在与其"行业及政府合作伙伴"展开合作。

英文来源：

OpenAI is launching Daybreak, an AI initiative focused on detecting and patching vulnerabilities before attackers find them. Daybreak uses the Codex Security AI agent that launched in March to create a threat model based on an organization’s code and focus on possible attack paths, validate likely vulnerabilities, and then automate the detection of the higher risk ones.
OpenAI just released its answer to Claude Mythos
OpenAI’s Daybreak combines GPT-5.5-Cyber and Codex Security.
OpenAI’s Daybreak combines GPT-5.5-Cyber and Codex Security.
Its launch comes just over a month after rival Anthropic announced Claude Mythos, a security-focused AI model it claimed was too dangerous to publicly release and only shared privately as a part of its own initiative, dubbed Project Glasswing. Still, that didn’t stop at least a few unauthorized parties from getting access.
However, OpenAI has so far lacked a similar security product. Like Glasswing, Daybreak isn’t built on just one AI model — OpenAI says “Daybreak brings together the most capable OpenAI models, Codex, and our security partners.”
Daybreak also involves specialized cyber models, including GPT-5.5 with Trusted Access for Cyber and GPT-5.5-Cyber, which began rolling out last week. OpenAI also says it’s working with its “industry and government partners” while it prepares to “deploy increasingly more cyber-capable models.”