OpenAI released a new cybersecurity model called ChatGPT Cyber on 6 November 2026, giving advanced tools to defenders and researchers in its Daybreak Red program. According to TechCrunch, the launch arrives as experts debate whether frontier labs are moving fast enough to protect digital infrastructure from AI-powered threats.
Key takeaways
- Max Justicz used an older OpenAI model to uncover a severe server vulnerability in Ubuntu code within two hours.
- OpenAI announced ChatGPT Cyber for high-tier users in its Daybreak Red program for vulnerability hunting.
- The AI giant stated that its models found 948 potential issues in the Linux kernel and 81 novel vulnerabilities in Firefox.
- OpenAI also launched a $1 billion initiative to support cybersecurity professionals protecting critical infrastructure.
What can the new OpenAI cybersecurity model do?
The newly announced ChatGPT Cyber model is directed at the highest tier of users in OpenAI's Daybreak Red program, providing offensive security researchers with tools for vulnerability hunting and development, according to TechCrunch. This follows tests where OpenAI used the model to scan the Linux kernel and prepare patches for 948 potential issues, alongside finding 81 novel vulnerabilities in the Firefox web browser.
Max Justicz, the co-founder and chief technology officer of Charge Robotics, previously demonstrated the power of these systems. TechCrunch reports that he used GPT 5.6 Sol through the Daybreak Blue program to find a convoluted server bug in Ubuntu software in just two hours by simply pointing ChatGPT at the codebase. He spent about $5,000 on tokens and subscriptions to find several exploits, reporting the Ubuntu issue to the Debian Project so they could issue a patch.
How does the Daybreak program work?
The Daybreak initiative gives defenders access to frontier models, though it faces mixed reactions from the community. According to TechCrunch, Justicz pointed out that OpenAI recently limited the blue tier to reduce risks, which leaves some professionals feeling restricted compared to bad actors using open-weight models. Only thousands of professionals currently participate in Daybreak, while Eric Wallace from OpenAI's technical staff noted that the company is dedicating substantial compute to pro bono open-source patching.
Why this matters
Digital infrastructure providers face a narrow window of time to secure systems before hackers gain easier access to advanced hacking methods. TechCrunch reports that while OpenAI initiated a $1 billion program for professionals working on water systems or the electrical grid, the Linux bug hunt cost $1.2 million. That figure remains lower than the approximately $6 million spent on solving a fluid dynamics problem, leaving some to question if labs are putting enough resources toward software safety.
Frequently asked questions
The new model is called ChatGPT Cyber, and it is designed to assist with vulnerability hunting and security development.
Max Justicz, the chief technology officer and co-founder of Charge Robotics, used an OpenAI model to find the vulnerability.
OpenAI models were used to scan and prepare patches for vulnerabilities in the Linux kernel, finding 948 potential issues.
Daybreak is an OpenAI program that gives cybersecurity defenders and researchers access to powerful frontier models for vulnerability hunting and development.
OpenAI announced a $1 billion initiative to support cybersecurity professionals in the Daybreak program who work on critical infrastructure.
Based on reporting by TechCrunch. Read the original article
