OpenAI rolls out new model for cybersecurity teams a month after Anthropic’s Mythos debuts

OpenAI introduces GPT‑5.5‑Cyber following outcry over anthropic myth
This is a modest upgrade focused on permissive cybersecurity tasks such as vulnerability triage and malware scanning.
Access is limited to vetted teams in the Trusted Access for Cyber (TAC) program, unlike Anthropic’s more restricted Mythos Preview.

OpenAI has introduced GPT-5.5-Cyber, an improved cybersecurity model aimed at taking some of the shine off Anthropic’s Mythos Preview.

Coming less than a month after the launch of GPT-5.4-Cyber, it is by no means a major upgrade and users should not expect many changes, OpenAI explained.

instead, users should expect a trained model to be more permissive when it comes to cybersecurity tasks, making it easier to use for tasks like vulnerability identification, triage, patch validation, and malware analysis.

Compete with the myth

“GPT-5.5-Cyber allows a smaller number of partners to study advanced workflows where specialized access behavior may be important,” OpenAI said in a blog post.

“The cyber defense ecosystem is large, and GPT-5.5 and GPT-5.5-Cyber play different roles to meet the needs of organizations and researchers, depending on the task, context, and safeguards for how the model is used. For most teams, GPT-5.5 with TAC is our most useful model for legitimate defensive work, with strong safeguards against misuse.”

As with the previous version, this edition will only be distributed to selected cybersecurity teams. However, unlike its main competitor, Mythos, which has only been offered to a handful of companies, OpenAI’s model will be offered to a broader set of users, members of the Trusted Access for Cyber (TAC) program.

At the time it introduced GPT-5.4-Cyber, OpenAI said it was expanding TAC to “thousands of verified individual defenders and hundreds of teams responsible for defending critical software.”

Anthropic first disclosed Project Glasswing in early April 2026, saying the Mythos Preview AI model was too powerful to give away for free. Apparently, he was able to reveal decades-old vulnerabilities in some of the most widely used operating systems and chain them together to create working exploits.