OpenAI has introduced a cybersecurity-focused model called ChatGPT Cyber for the highest tier of its Daybreak program, an acknowledgment that frontier AI is no longer just a productivity toy. It is already being used to find bugs in software that large parts of the internet run on.
The useful number here is not a valuation or a user count. It is 948: the number of potential Linux kernel issues OpenAI said the model helped identify and prepare patches for. The company also said it found 81 “novel” vulnerabilities in Firefox. Those figures suggest real scanning power, though “potential issues” is not the same thing as confirmed exploitable flaws, and OpenAI has not turned that claim into a public accounting of severity.
What makes the launch harder to dismiss is the example TechCrunch tied to it. Max Justicz, cofounder and CTO of Charge Robotics, told the publication he used OpenAI’s GPT 5.6 Sol through Daybreak Blue to find a severe vulnerability in Ubuntu-related software maintained by the Debian Project. Debian issued a patch, and TechCrunch reported there was no known evidence of exploitation. That is the reassuring part. The less reassuring part is how he described the process: point the model at code that had already produced one bug in the past, then wait a couple of hours.
OpenAI’s answer is to widen access for defenders, at least selectively. ChatGPT Cyber is being made available to Daybreak Red, aimed at offensive security researchers working on vulnerability hunting and development. The company also announced a $1 billion initiative for cybersecurity professionals in the program who work on critical infrastructure.
That sounds large until it meets the denominator. OpenAI said only “thousands” of cybersecurity professionals are in Daybreak, against a much larger pool of possible attackers. Justicz also told TechCrunch he spent about $5,000 on subscriptions and tokens to find several exploits, which hints at another constraint: these tools are getting better, but useful access is not free and apparently not broad.
OpenAI says it is dedicating a significant amount of frontier compute to finding and patching open-source vulnerabilities at no charge. TechCrunch reported the Linux bug-hunting effort cost about $1.2 million in token spend. Cheap compared with some grand AI research projects, perhaps. Expensive if the industry still treats patching the plumbing as a side quest.
Isla Camilleri 
Alex de Valletta 
Ryan C