As AI-led attacks multiply, OpenAI launches a new cyber model

AI agents are increasingly being used to conduct cyberattacks, including compromising websites and creating fake profiles to infiltrate systems. In response, OpenAI has expanded Daybreak, its cyber defense service, by introducing two service tiers: Blue for basic defensive work like incident response and malware analysis, and Red for more advanced security testing and vulnerability research. OpenAI also released a new specialized model called GPT-5.6-Cyber, available only through the Red tier to approved customers. The development reflects both the growing threat from AI-led attacks and the business opportunity for AI labs to sell protection against the security risks they understand firsthand.