Initial reports suggested this was a new AI model, drawing comparisons to Anthropic, which recently limited access to its “Mythos Preview” model to select tech and security firms due to its advanced hacking-related capabilities.

However, Axios has since clarified that the restricted rollout applies only to the cybersecurity product — not to OpenAI’s upcoming “Spud” model.

The rollout is expected to take place through OpenAI’s “Trusted Access for Cyber” pilot program, launched in February alongside the release of GPT-5.3 Codex. Participants in the program are granted access to highly capable models designed for defensive security work, along with up to $10 million in API credits.