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

This move comes after OpenAI publicly criticised Anthropic for restricting access to Mythos, accusing the rival lab of gatekeeping advanced AI capabilities.

The restrictions on GPT-5.5 Cyber mean that only a select group of cybersecurity professionals deemed 'critical cyber defenders' will receive initial access. OpenAI has not specified how these defenders will be vetted or what criteria will be used. Anthropic had faced similar criticism for limiting Mythos to a curated list of partners.

The announcement highlights a broader tension in the AI industry between the push for open access and the perceived need to control powerful AI models. Critics argue that such restrictions create a two-tier system, while companies maintain that careful stewardship is necessary to prevent misuse. Neither OpenAI nor Anthropic has indicated plans to expand access beyond the initial groups.