OpenAI's ChatGPT Cyber opens to top Daybreak Red tier

OpenAI's move comes as outside researchers are already using its models to find serious flaws. Max Justicz, co-founder and CTO of Charge Robotics, used GPT 5.6 Sol through Daybreak Blue to find a severe vulnerability in servers running Ubuntu, which the Debian Project patched. Justicz said he spent about $5,000 on subscriptions and tokens to find several exploits, and warned that less-guarded models could enable "extremely crazy bugs" to be found and exploited over the next few months.

The new release also highlights a tension in OpenAI's approach. Justicz noted that the lab appears to have recently "nerfed" Daybreak Blue, limiting its usefulness for vulnerability hunting—a step that may restrict bad actors but also leaves defenders at a disadvantage if attackers can use open-weight models with abliterated guardrails. OpenAI says just "thousands" of cybersecurity professionals are in Daybreak, versus a far larger number of potential hackers.

OpenAI researcher Eric Wallace said the company is allocating a significant amount of frontier compute to discovering and patching vulnerabilities in open source codebases, fully pro bono. The Linux bug hunt cost $1.2 million, which is significantly less than the approximate $6 million token spend OpenAI invested in a fluid dynamics problem. If there is a narrow window to patch digital infrastructure, more urgency may be required.