Singapore – Check Point Software Technologies has expanded its artificial intelligence partnership with OpenAI, integrating the latter’s Daybreak frontier cyber models across its security platform to help organisations detect, validate, and remediate threats faster.
The move builds on a collaboration the two companies launched three months ago through Daybreak, OpenAI’s cyber defence initiative, according to Check Point.
Jonathan Zanger, Chief Technology Officer at Check Point, said the partnership has since progressed further, with OpenAI’s frontier cyber models now embedded into Check Point products and workflows through the Daybreak Defence Network.
Last week, Check Point joined more than 100 technology and security companies in backing OpenAI’s call for a coordinated, global surge in cyber defence, the company said.
According to Check Point, security must continue to evolve as new threats emerge, software stacks change, and organisations find new ways to put AI to work.
As attackers move faster and operate with greater scale and sophistication, the company said defenders need equivalent advances in AI to keep pace — a principle it described as central to its collaboration with OpenAI.
Check Point said the integration is designed to help it build, test and continuously strengthen security, while embedding these capabilities directly into the workflows customers already rely on.
Per the company, OpenAI’s frontier cyber reasoning is becoming part of how Check Point discovers risk, validates what is actually exploitable, accelerates protection, investigates threats, and secures customer environments.
Combined with Check Point’s existing security intelligence and enforcement capabilities, this is intended to help customers move from large volumes of security data toward validated risk, actionable decisions and faster protection, the company said.
The rollout spans four areas of Check Point’s platform, based on details shared by the company.
Under Agentic Exposure Validation, part of its Exposure Management offering, Check Point is piloting OpenAI’s frontier cyber models within its multi-agent pipeline, aiming to separate genuine, exploitable risk from theoretical findings. This pairs OpenAI’s reasoning capabilities with Check Point’s own security context, controls and evidence to validate attack paths, prioritise proven risk and speed up remediation, according to the company.
Within Keystone, its agentic security management system, Check Point said it is bringing OpenAI’s cyber reasoning into an autonomous, intent-driven approach to network security. Rather than requiring manual policy management, customers define their intent while AI continuously assesses the environment, identifies risk and adjusts controls accordingly, the company said.
Check Point’s Autonomous Workspace Platform, which correlates telemetry from email, endpoint, mobile, and browser sources into investigated incidents, is also being enhanced with OpenAI’s reasoning capabilities. This is intended to help investigate complex threats spanning malware behaviour, attacker techniques and credential-abuse chains, while delivering clearer verdicts and reducing the workload on security teams, according to Check Point.
The fourth area, NexPloit, converts vulnerability data into verified attack material to support automated protection development without relying on publicly available exploit code, the company said. Check Point said it is using OpenAI’s frontier cyber models to accelerate the process of understanding vulnerable code and patches, identifying realistic exploitation paths, and reaching verified results more quickly — intended to translate into faster protection against newly disclosed vulnerabilities.
Through the Daybreak Defence Network, Check Point said it aims to apply frontier AI across the security lifecycle to help customers identify proven risk, act faster, automate more of their operations and prevent more attacks.
Check Point described its rollout as a phased approach. Some capabilities are already in production, the company said, while others remain in development and testing with design partners, with further releases expected as the technology matures.
According to Check Point, every deployment follows the same governance discipline: controlling what the model can access, constraining what it can act on, testing and verifying its output, and only then expanding its role within approved security workflows.
The company said its longer-term goal is to translate advances in frontier models into security outcomes that customers can trust, while maintaining the controls required for enterprise-grade security.
Check Point framed its work with OpenAI as operating on two fronts — using frontier AI to strengthen its own defences, and helping customers adopt OpenAI’s technologies securely. Both dimensions are becoming more important, the company said, as AI moves beyond answering questions to writing code, running enterprise agents and taking autonomous action.
Security must keep pace with these shifts, according to Check Point, citing evolving threat actor capabilities, changing technology environments and customers’ expanding use of AI. For Check Point, that means continuously applying AI advances to improve threat prevention, accelerate security operations and allow systems to handle more complexity independently.
Check Point said its broader objective is to give organisations the confidence to adopt AI while remaining protected against emerging risks, positioning its OpenAI partnership as a way to put frontier AI to work for defenders and their customers alike.
