Check Point Software Technologies has expanded its integration of OpenAI’s frontier cyber defence models across its security platform, bringing the models into products spanning exposure management, network security, incident investigation and vulnerability research.

The move builds on Check Point and OpenAI’s collaboration through Daybreak, OpenAI’s cyber defence initiative, which the two companies began expanding three months ago. Through the Daybreak Defense Network, Check Point has now brought OpenAI’s frontier cyber models into its security workflows, and last week joined more than 100 technology and security companies backing OpenAI’s call for a collective global surge in cyber defence.

Frontier AI applied across the security lifecycle

In an announcement piece, Jonathan Zanger, Chief Technology Officer at Check Point, outlined four areas where customers will benefit. In Agentic Exposure Validation, OpenAI’s frontier cyber models are being piloted to combine advanced reasoning with Check Point’s security context to validate attack paths and prioritise proven risk. In Keystone, an intent-driven approach to network security management, the models help investigate potential attack paths and identify appropriate remediation.

The Autonomous Workspace Platform applies OpenAI’s cyber reasoning to investigate complex threats across malware behaviour, attacker techniques and credential-abuse chains, while NexPloit uses the models to accelerate vulnerability research by understanding vulnerable code and patches and identifying realistic exploitation paths — without relying on publicly available exploit code.

“Putting frontier AI to work to help defenders stay ahead” is central to the collaboration, wrote Jonathan Zanger, Chief Technology Officer, Check Point Software Technologies.

He added that the two companies are taking a phased approach to the rollout, with some capabilities already in production, others in development with design partners, and further releases planned as governance controls mature around what the models can see, act on and verify before taking on more autonomous work.

No regional rollout details disclosed

The announcement was framed globally, with no Southeast Asia-specific rollout details or customer references disclosed at this stage.