OpenAI reportedly following Anthropic's lead in restricting access to powerful cybersecurity AI

OpenAI is also working on a new AI model with advanced cybersecurity capabilities that will only be available to a small group of companies, according to Axios. The approach mirrors Anthropic's move on Tuesday, when it restricted access to its new Mythos Preview model to select tech and security firms because of its powerful hacking capabilities.

Back in February, OpenAI had already launched its "Trusted Access for Cyber" pilot program after releasing GPT-5.3-Codex, its most capable cybersecurity model to date. Participants get access to particularly powerful models for defensive security work, backed by $10 million in API credits. Whether OpenAI will eventually make the new model more widely available remains unclear. Anthropic has ruled out a public release of Mythos Preview, saying models in the Mythos class won't ship until adequate safety guardrails are in place.