OpenAI debuts GPT-6 Cyber for vetted security teams
OpenAI has introduced GPT-6 Cyber, a model restricted to vetted companies and security teams that identified 81 new vulnerabilities in Firefox during testing. The launch includes new identity safeguards and a partnership with Dragos, positioning the model as a tool for enterprise security rather than general use.
OpenAI has debuted GPT-6 Cyber, a model restricted to vetted companies and Daybreak Red security teams. During testing, the model identified 81 new vulnerabilities in Firefox and developed and validated a critical remote code execution bug in the Bratley compression library. To prevent misuse by hostile actors in North Korea or China, OpenAI is implementing new identity checks and safeguards. The company also announced a partnership with operational technology security firm Dragos.
The move is seen as a strategic effort to drive enterprise revenue by positioning OpenAI as a solution to the security risks highlighted by competitors like Anthropic. One analyst argued that OpenAI is strategically using the model's dangerous capabilities to generate enterprise revenue, contrasting it with Anthropic's approach of emphasizing model danger. "If you want to make a lot of money, the best thing to do is scare everyone to saying your model's going to hack the world and then say, oh, look, we have OpenAI, GPT-6, Cyber, and you can go fix all the problems that our unreleased model could find," the analyst said.