- OpenAI announced GPT-6 Cyber for advanced cybersecurity research and vulnerability discovery.
- The model discovered 81 high-severity flaws in Firefox and built exploits for Linux.
- OpenAI restricts access to a small group of approved cybersecurity teams.
- Safety evaluations showed 100% compliance when explicit instructions were provided.
In this guide
What is GPT-6 Cyber?
GPT-6 Cyber is an artificial intelligence model created by OpenAI specifically for cybersecurity research and vulnerability detection. It audits software code to locate security flaws and generate fixes before attackers can exploit them.
- Developer
- OpenAI
- Announcement Date
- 5 November 2026
- Safety Framework Classification
- Critical
- Availability
- Restricted to approved security teams
- Primary Capabilities
- Vulnerability discovery, patch creation, code review
- Predecessor Model
- GPT-5.6 Cyber
OpenAI announced GPT-6 Cyber on 5 November 2026 after using it internally to power security research. OpenAI reports that GPT-6 Cyber significantly outperforms the previous GPT-5.6 Cyber model at identifying and repairing software bugs.
What vulnerabilities has GPT-6 Cyber discovered?
GPT-6 Cyber has demonstrated the ability to discover high-severity software flaws across web browsers and operating systems. During security research campaigns, GPT-6 Cyber identified complex vulnerabilities that had evaded earlier automated security models.
In testing on the Firefox web browser, GPT-6 Cyber uncovered 81 new security defects that Mozilla classified as high severity. GPT-6 Cyber submitted proposed patches for these defects and established that 30 previously known bugs posed a higher risk than originally categorized. Mozilla subsequently issued updates containing these fixes to millions of users.
GPT-6 Cyber also generated working exploits for 948 Linux kernel vulnerabilities in controlled research environments. In comparison, OpenAI stated that its older GPT-5.6 Sol model produced 166 working exploits under the same evaluation conditions. OpenAI shared 304 critical findings and suggested patches directly with Linux maintainers.
Who can access GPT-6 Cyber?
GPT-6 Cyber is not available to the public or general developers. OpenAI restricts access to GPT-6 Cyber to a small group of verified cybersecurity organizations and internal research groups.
OpenAI assigns GPT-6 Cyber a Critical rating within its cybersecurity safety framework due to its advanced exploitation capabilities. Commercial technology firms, including Cisco and CodeRabbit, currently use GPT-6 Cyber to conduct automated code reviews.
Most users cannot access GPT-6 Cyber directly. Everyday consumers benefit indirectly when software vendors deploy patches created with assistance from GPT-6 Cyber.
How safe is GPT-6 Cyber?
GPT-6 Cyber features specialized alignment training designed to keep automated testing within designated network boundaries. OpenAI evaluated GPT-6 Cyber in closed testing environments surrounded by vulnerable target systems that were explicitly outside the scope of the assigned task.
In trials conducted without safety instructions, GPT-6 Cyber refrained from touching out-of-scope services in 99.93% of tests. When explicit safety instructions were included, GPT-6 Cyber maintained perfect compliance and avoided restricted systems in 100% of test runs.
OpenAI also tested GPT-6 Cyber on its own server infrastructure. Controlled agents powered by GPT-6 Cyber successfully broke out of isolated container environments and executed code on underlying machines, allowing OpenAI engineers to patch system flaws and tighten internal network restrictions.
Frequently asked questions
What is GPT-6 Cyber?
Can the public access GPT-6 Cyber?
Which security flaws has GPT-6 Cyber uncovered?
Which companies currently use GPT-6 Cyber?
How safe is GPT-6 Cyber during security testing?
Sources: Digital Trends
Want your brand to be the answer AI gives?
See how ready your website is for ChatGPT, Gemini and Perplexity, free, in about ten seconds.