OpenAI Debuts GPT-6 Cyber, Enhancing Cybersecurity Performance and Safety Guardrails
OpenAI has launched GPT-6 Cyber, a new cybersecurity-focused AI model designed to improve vulnerability discovery, patch development, and adherence to safety guardrails. According to OpenAI officials, GPT-6 Cyber has found at least 81 novel vulnerabilities in the Firefox web browser and developed exploits for 948 Linux kernel vulnerabilities. The model also successfully developed and validated a critical remote code execution vulnerability in Brotli, an open-source data compression tool. These vulnerabilities were reported and patched. GPT-6 Cyber is initially available to a limited number of customers in OpenAI’s Daybreak Red program, which provides access to advanced cybersecurity models for offensive red teaming. The company emphasizes that the new model includes enhanced safeguards to prevent misuse and ensure it operates within authorized parameters.
