OpenAI is preparing to launch a preview version of its new AI model, GPT-6 Cyber, within the coming weeks. The model is designed to assist clients in automating workflows and identifying vulnerabilities to counter AI-driven cyberattacks.
The company will also introduce a new tool to facilitate the secure and automated deployment of the model. This initiative is part of OpenAI's strategy to provide specialized security solutions, with the model already undergoing alpha testing by a select group of clients under the Daybreak Red program.
GPT-6 Cyber marks the fourth security-focused model released by OpenAI this year, following the launches of GPT-5.4, GPT-5.5, and GPT-5.6 Cyber. The company's security division is led by Chief Commercial Officer Dali Rajic, who joined in August.
OpenAI has committed $1 billion to subsidize the use of its security products for critical infrastructure. Additionally, the company plans to unveil approximately a dozen new products at the DevDay conference on September 29, 2026.
