OpenAI said GPT-6 Cyber identified 81 previously undisclosed Firefox vulnerabilities and supplied Mozilla with suggested fixes, showing how AI tools can contribute to browser security work. Mozilla has begun distributing security updates.
OpenAI’s Critical rating concerns GPT-6 Cyber’s cybersecurity capabilities under its Preparedness Framework. Mozilla rated the reported browser flaws high severity; that assessment concerns the vulnerabilities, not the model.
OpenAI also said the model reproduced 1,724 Linux kernel vulnerabilities and developed 948 working privilege-escalation exploits. The company says GPT-6 Cyber was trained to locate and verify exploitable software weaknesses within authorized testing boundaries. OpenAI said its researchers prepared evidence and repeatable tests supporting the Firefox findings. EconoTimes previously covered the Linux exploit findings.
OpenAI says approved Daybreak Red partners can first access GPT-6 Cyber through its API and Codex. OpenAI plans to extend access to all Daybreak Red customers later, but has not announced a date for that rollout.
