OpenAI unveils GPT-6 Cyber for vulnerability detection and exploitation

OpenAI has officially launched GPT-6 Cyber, a powerful neural network designed to identify software vulnerabilities and assist in creating exploits. The company has already successfully used this model to discover hundreds of previously unknown security flaws in Firefox and Linux.

• GPT-6 Cyber is rated as a Critical risk model, surpassing the previous GPT-5.6 Cyber version.
• The AI discovered 81 vulnerabilities in Firefox and developed exploits for 948 Linux kernel vulnerabilities.
• Access is restricted to vetted participants of the Daybreak Red security program.
• Developers were provided with data on 588 previously unknown Linux vulnerabilities along with proposed fixes.

Model Capabilities and Impact GPT-6 Cyber represents a significant leap in automated security analysis. By detecting 81 vulnerabilities in the Firefox browser, the AI has prompted Mozilla to initiate critical security patches. Even more impressive, the model generated exploits for 948 Linux kernel vulnerabilities, 588 of which were previously undisclosed to the public and developers. OpenAI has shared these findings and recommended patches to the Linux developer community. Because of its ability to generate functional exploits, OpenAI has classified this model as Critical. The company is taking a cautious approach to deployment to prevent the tool from being misused by malicious actors.

Access Control and Security Strategy Due to the potent nature of the model, access remains strictly limited. Only participants in the Daybreak Red program, who are involved in authorized security testing, can currently utilize GPT-6 Cyber. OpenAI aims to expand this access to more security professionals in the coming weeks while maintaining rigorous oversight. The core mission of this project is to provide security experts with the tools to find and patch vulnerabilities faster than attackers can exploit them. By accelerating the discovery phase of the security lifecycle, OpenAI hopes to significantly harden critical software infrastructure.

Implications for the Tech Industry The introduction of GPT-6 Cyber changes the landscape of cybersecurity, forcing a shift toward AI-assisted defense. For developers and cybersecurity professionals, this signals that automated vulnerability assessment will become the new standard. This technology is poised to become an essential tool for organizations tasked with protecting critical infrastructure from increasingly sophisticated cyber threats.