After dissing Anthropic for limiting Mythos, OpenAI restricts access to Cyber, too

After publicly dissing Anthropic for locking down its cybersecurity tool Mythos to only select users, OpenAI CEO Sam Altman confirmed that his own company would do the same with GPT-5.5 Cyber. In an X post, Altman announced that Cyber would roll out 'to critical cyber defenders' in the coming days, with access granted through an application process that requires credentials and planned use details. Cyber is a specialized toolkit designed for penetration testing, vulnerability identification and exploitation, and malware reverse engineering—essentially helping companies find and fix security holes.

The move is a stark reversal for Altman, who had previously called Anthropic's restriction of Mythos 'fear-based marketing.' Critics quickly called out the double standard, especially after an unauthorized group reportedly managed to access Mythos anyway. OpenAI says it's working to make Cyber more widely available by coordinating with the U.S. government and identifying more legitimate cybersecurity professionals. The irony is not lost on industry watchers: a company that mocked exclusivity is now embracing it, raising questions about how to safely distribute powerful offensive security tools.