<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title><![CDATA[OpenAI 推出 GPT-5.5-Cyber，限定”关键网络防御者”访问，与 Anthropic Mythos 形成同台竞争]]></title><description><![CDATA[<p dir="auto">OpenAI CEO Sam Altman 4 月 30 日在 X 宣布，将于数日内向”关键网络防御者（critical cyber defenders）“开始分发 GPT-5.5-Cyber——一款基于 GPT-5.5 调优的网络安全专用前沿模型，称将”与整个生态系统及政府合作确定网络安全领域的可信访问机制，希望快速帮助保护企业与关键基础设施”。OpenAI 未公布任何技术参数或性能基准；已知的是，母版 GPT-5.5 在 OpenAI 自有”准备框架”的网络安全领域获评”高（High）“级，处于”关键（Critical）“门槛之下——后者的定义是可在无人协助情况下自主开发零日漏洞利用代码。访问渠道延续此前的”网络安全可信访问（TAC）“计划，覆盖对象预计比 Anthropic Mythos 的约 50 家机构更广，拟纳入政府机构、关键基础设施运营商、安全厂商、云平台与金融机构。前代 GPT-5.4-Cyber 已通过 TAC 向数千名安全从业者发放，并配套 1000 万美元 API 额度资助。</p>
<p dir="auto">此次发布的直接背景是 Anthropic 4 月 7 日发布的 Claude Mythos Preview——后者被披露可在主流操作系统与浏览器中自主发现数千个零日漏洞，并能将多个漏洞链式组合形成复杂攻击，Anthropic 随即将其限制在约 50 家组织内并发动白宫级别的政策角力。GPT-5.5-Cyber 被业界解读为 OpenAI 对这一节点的直接回应：两家公司正在用相近的”受控分发”逻辑处理各自最具网络能力的模型，但 OpenAI 选择更宽泛的覆盖面、更低调的发布方式，以及”帮助防御者”而非”发现漏洞”的叙事口径。</p>
<p dir="auto"><a href="https://www.theverge.com/ai-artificial-intelligence/921073/openai-sam-altman-new-cybersecurity-model-gpt-5-5-cyber" target="_blank" rel="noopener noreferrer nofollow ugc">The Verge</a> | <a href="https://techbriefly.com/2026/04/30/new-gpt-5-5-cyber-model-follows-anthropics-secret-mythos-release/" target="_blank" rel="noopener noreferrer nofollow ugc">TechBriefly</a> | <a href="https://www.businesstoday.in/technology/story/openai-to-roll-out-gpt-55-cyber-with-restricted-access-sam-altman-528193-2026-04-30" target="_blank" rel="noopener noreferrer nofollow ugc">BusinessToday</a></p>
<p dir="auto"><div class="card col-md-9 col-lg-6 position-relative link-preview p-0">



<a href="https://www.theverge.com/ai-artificial-intelligence/921073/openai-sam-altman-new-cybersecurity-model-gpt-5-5-cyber" title="OpenAI’s new security model is for ‘critical cyber defenders’ only">
<img src="https://platform.theverge.com/wp-content/uploads/sites/2/2026/04/STK155_OPEN_AI_CVirginia_C.jpg?quality=90&strip=all&crop=0%2C10.732984293194%2C100%2C78.534031413613&w=1200" class="card-img-top not-responsive" style="max-height: 15rem;" alt="Link Preview Image" />
</a>



<div class="card-body">
<h5 class="card-title">
<a class="text-decoration-none" href="https://www.theverge.com/ai-artificial-intelligence/921073/openai-sam-altman-new-cybersecurity-model-gpt-5-5-cyber">
OpenAI’s new security model is for ‘critical cyber defenders’ only
</a>
</h5>
<p class="card-text line-clamp-3">Like Anthropic’s Mythos, GPT-5.5-Cyber will first be released to ‘trusted’ entities. </p>
</div>
<a href="https://www.theverge.com/ai-artificial-intelligence/921073/openai-sam-altman-new-cybersecurity-model-gpt-5-5-cyber" class="card-footer text-body-secondary small d-flex gap-2 align-items-center lh-2">



<img src="https://www.theverge.com/static-assets/icons/favicon-32x32.png" alt="favicon" class="not-responsive overflow-hiddden" style="max-width: 21px; max-height: 21px;" />















<p class="d-inline-block text-truncate mb-0">The Verge <span class="text-secondary">(www.theverge.com)</span></p>
</a>
</div></p>
<p dir="auto"><div class="card col-md-9 col-lg-6 position-relative link-preview p-0">

<div class="card-body">
<h5 class="card-title">
<a class="text-decoration-none" href="https://techbriefly.com/2026/04/30/new-gpt-5-5-cyber-model-follows-anthropics-secret-mythos-release/">
Just a moment...
</a>
</h5>
<p class="card-text line-clamp-3"></p>
</div>
<a href="https://techbriefly.com/2026/04/30/new-gpt-5-5-cyber-model-follows-anthropics-secret-mythos-release/" class="card-footer text-body-secondary small d-flex gap-2 align-items-center lh-2">



<img src="https://techbriefly.com/favicon.ico" alt="favicon" class="not-responsive overflow-hiddden" style="max-width: 21px; max-height: 21px;" />



<p class="d-inline-block text-truncate mb-0"> <span class="text-secondary">(techbriefly.com)</span></p>
</a>
</div></p>
]]></description><link>https://welinux.com//topic/132/openai-推出-gpt-5.5-cyber-限定-关键网络防御者-访问-与-anthropic-mythos-形成同台竞争</link><generator>RSS for Node</generator><lastBuildDate>Sat, 02 May 2026 18:06:04 GMT</lastBuildDate><atom:link href="https://welinux.com//topic/132.rss" rel="self" type="application/rss+xml"/><pubDate>Fri, 01 May 2026 08:19:58 GMT</pubDate><ttl>60</ttl></channel></rss>