Anthropic 將 Mythos 安全防線擴展至 15+ 個國家
anthropic 將 project glasswing 擴展到 15+ 個國家的 150 個組織，利用 claude mythos 模型查找關鍵基礎設施中的零日漏洞。
anthropic scaled project glasswing to about 150 new organizations across more than 15 countries, providing access to claude mythos to secure codebases for power, water, and healthcare sectors. partners include nato, the eu's enisa, okta, samsung, and sk hynix, chosen because a successful attack on their systems would affect over 100 million people. the move follows a confidential ipo filing and comes as openai released its own cybersecurity model gpt-5.5-cyber to partners.