OpenAI released a new cybersecurity model. Is that enough?

OpenAI has announced ChatGPT Cyber, a security-focused model available to the top tier of its Daybreak program, Daybreak Red, which serves offensive researchers hunting for vulnerabilities. The company said the model scanned the Linux kernel and prepared patches, flagging 948 potential issues, and also turned up 81 "novel" flaws in Firefox. OpenAI says it wants to widen Daybreak to organizations of all sizes, independent researchers and open-source maintainers, and has announced a $1 billion initiative for participants protecting critical infrastructure such as power grids and water systems.

TechCrunch frames the launch against the experience of Max Justicz, CTO and co-founder of Charge Robotics and a hobbyist in security. He told the outlet that GPT 5.6 Sol, accessed through the Daybreak Blue defender tier, found a severe Ubuntu server takeover flaw within two hours of being pointed at code where he had previously found a bug. He reported it to the Debian Project, which issued a patch, and there is no known evidence of exploitation. Justicz spent about $5,000 finding several exploits and predicts serious bugs will be found and exploited in coming months.

The article questions whether the effort matches the risk. Justicz says Daybreak Blue appears to have been recently weakened, though he joined simply by uploading a driver's license photo. Security professionals complain that attackers can use open-weight models stripped of safeguards, while OpenAI counts only thousands of program participants. OpenAI researcher Eric Wallace says significant compute is going to pro bono patching, but TechCrunch notes the Linux hunt cost $1.2 million, against roughly $6 million in tokens spent on a fluid dynamics problem.