OpenAI Releases GPT-6 Cyber, Flags Critical Security Risks
OpenAI has released GPT-6 Cyber, a new artificial intelligence model designed to identify and exploit software security vulnerabilities, to a limited group of vetted customers, classifying it as “Critical” under its cybersecurity safety framework.
The model is designed to help security researchers discover vulnerabilities, assess how attackers could exploit them and develop fixes before malicious actors can take advantage of weaknesses.
OpenAI said the release comes amid growing attempts by cybercriminals and other malicious groups to misuse artificial intelligence for cyberattacks.
According to the company, GPT-6 Cyber has already demonstrated significant capabilities in identifying software flaws.
OpenAI researchers used the model to discover 81 previously unknown vulnerabilities in Mozilla Firefox, which Mozilla classified as high severity.
The browser developer has begun releasing security updates to address the issues.
Also Read: Qualcomm, SpaceX Sign Major 5G Deal for Starlink Mobile Network
The model also helped develop working exploits for 948 Linux kernel vulnerabilities, including 588 that were previously unknown, OpenAI said. In addition, the company has disclosed 304 critical Linux security findings to maintainers and provided proposed patches to help address them.
Also Read: Brent Crude Falls 10% Below $100 on Iran War Peace Hopes
OpenAI has raised the model's cybersecurity risk classification to “Critical”, up from “High” for its previous GPT-5.6 Cyber model, reflecting its increased ability to identify and develop exploits for security weaknesses.
The company said it is also observing a rise in attempts to misuse its AI systems, including activities by malicious cyber actors and groups attempting to distil its models.
GPT-6 Cyber will initially be available only to selected customers through Daybreak Red, OpenAI's restricted access programme for advanced and authorised cybersecurity activities, including penetration testing and exploit development.
Also Read: Hegseth Approves More Armed National Guard Patrols in DC
The limited rollout is intended to give trusted security professionals access to the model's advanced capabilities while reducing the risk of its misuse.