<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>GPT-5.5-Cyber on AI内参</title>
    <link>https://www.neican.ai/tags/gpt-5.5-cyber/</link>
    <description>Recent content in GPT-5.5-Cyber on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Tue, 23 Jun 2026 12:10:04 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/gpt-5.5-cyber/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>OpenAI放出&#34;安全完全体&#34;GPT-5.5-Cyber：修Bug比写代码还快，还顺手击败了Anthropic的Mythos 5</title>
      <link>https://www.neican.ai/insights/openaigpt-55-cyberbuganthropicmythos-5-20260623121004945-2/</link>
      <pubDate>Tue, 23 Jun 2026 12:10:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/openaigpt-55-cyberbuganthropicmythos-5-20260623121004945-2/</guid>
      <description>OpenAI发布了Daybreak安全计划，核心是GPT-5.5-Cyber完整版，在安全基准测试中击败Anthropic的Mythos 5，同时推出Codex Security插件和Patch the Planet开源修复计划，目标是从漏洞发现走向真正修复。但网友更期待GPT-5.6，安全专用模型显然不如通用迭代吸睛。</description>
    </item>
    <item>
      <title>OpenAI刚派GPT-5.5-Cyber上演“修补地球”大片，转头自家Codex就爆出“SSD杀手”Bug</title>
      <link>https://www.neican.ai/insights/openaigpt-55-cybercodexssdbug-20260623111004351-0/</link>
      <pubDate>Tue, 23 Jun 2026 11:10:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/openaigpt-55-cybercodexssdbug-20260623111004351-0/</guid>
      <description>OpenAI刚刚发布了号称能“修补地球”的最强网安模型GPT-5.5-Cyber，同时推出了Codex Security插件，目标是把AI变成全球网络安全的底层系统。然而，网友们发现Codex存在一个致命Bug，会以极高速率疯狂写入日志，一年就能写废一块固态硬盘，堪称“SSD杀手”。这波冰火两重天的操作，让人不得不感叹：技术的光芒有时也会闪瞎自己的眼。</description>
    </item>
  </channel>
</rss>
