OpenAI debuts new cybersecurity-focused AI model amid growing alarm about AI-driven cyber attacks
Reported by Fortune1 source ↓
OpenAI says GPT-6 Cyber is designed for advanced vulnerability research and exploit development. It will first go to a small group of approved Daybreak Red customers, with access due to expand over the coming weeks.
Key points
- GPT-6 Cyber will initially be available only to approved Daybreak Red customers
- OpenAI says the model found 81 unknown, high-severity Firefox vulnerabilities
- OpenAI says it developed exploits for 948 Linux kernel vulnerabilities, including 588 previously unknown flaws
- OpenAI is expanding Astra testing for defensive security through Daybreak Blue
Details
OpenAI says GPT-6 Cyber outperforms its general-purpose GPT-6 Astra model on most cyber tasks. The company says the model is intended for advanced vulnerability research and exploit development, work that can help defenders but could also be used in attacks.
In tests on Firefox and the Linux kernel, OpenAI says researchers used GPT-6 Cyber to find 81 previously unknown, high-severity Firefox vulnerabilities and develop working exploits for 948 Linux kernel vulnerabilities, including 588 previously unknown flaws. The Linux exploits could allow someone with ordinary local access to gain root privileges. OpenAI says researchers developed working exploits for more than five times as many Linux vulnerabilities as in a July study using GPT-5.6 Sol.
OpenAI offers trusted cybersecurity professionals access through Daybreak Red for advanced work such as attack simulation and tool development, and Daybreak Blue for defensive tasks. Separately, the company says it is expanding early testing of Astra for defensive security to more organizations and individual researchers through Daybreak Blue.
Written from coverage as of