OpenAI rolls out new GPT-5.5-Cyber to vetted cybersecurity teams

OpenAI on Thursday announced A variant of the latest AI model, GPT-5.5-Cyber, is being rolled out in a limited preview capacity to vetted cybersecurity teams, a month after rival Anthropic wowed investors and government officials with the Claude Mythos Preview.

The preview of GPT-5.5-Cyber ​​is not intended to be a major leap forward in terms of cyber capability, but is instead trained to be more lenient on security-related tasks, OpenAI said in a blog post. OpenAI announced GPT-5.5 late last month.

The company said that with the cyber-specific version, surveyed teams will have an easier time using OpenAI’s latest model for workflows such as vulnerability identification and prioritization, patch verification, and malware analysis. Overall, the security measures built into the current GPT-5.5 model would have made this even more difficult.

“GPT‑5.5‑Cyber ​​allows a smaller group of partners to examine advanced workflows where private access behavior may be important,” OpenAI said in the blog post.

Anthropic, which launched Mythos last month, has decided to limit access to a select group of companies as part of a new cybersecurity initiative called Project Glasswing. Anthropic CEO Dario Amodei met with senior members of the Trump administration to talk about the model and its potential power, despite the company being blacklisted by the Pentagon a few weeks ago.

Federal Reserve Chairman Jerome Powell and Treasury Secretary Scott Bessent met with major U.S. bank CEOs last month to discuss Mythos, and Vice Chairman J.D. Vance and Bessent met with leading technology CEOs ahead of the model’s launch.

WRISTWATCH: Federal agencies seek access to Mythos, leading to Antropik CEO meeting with White House