OpenAI details how NATO and European cyber agencies are using its models for defense
OpenAI says NATO's communications agency uses up to 300 billion tokens a month of its cybersecurity models through Daybreak, and that ENISA, CERT Polska and others have used them to find real vulnerabilities.
OpenAI has published an update on how government cyber defenders in NATO and across Europe are using its models through Daybreak, its program that gives vetted defenders access to advanced AI for authorized security work such as reviewing code, investigating suspicious activity, validating vulnerabilities and testing fixes.
NATO
OpenAI says it works with NATO's Communications and Information Agency (NCIA) to give NATO's core institutions access to its cybersecurity models. According to the company, NATO has used up to 300 billion raw input and output tokens per month as part of this work. NCIA's general manager, Dylan Browne, is quoted saying the agency will study token use and the economics of these models as it scales, to help it compare options and make procurement decisions.
OpenAI also supported NATO's 2026 Crossed Swords exercise, run by the NATO Cooperative Cyber Defence Centre of Excellence. A new training stream had participants build, deploy and evaluate AI agents for cyber operations using Daybreak, with an emphasis on keeping humans in charge.
Other European defenders
OpenAI says it has also given access to the EU cyber agency ENISA and to national agencies in France, Germany and Poland, as well as Ukraine's Ministry of Digital Transformation for protecting civilian infrastructure. The examples it cites:
- ENISA used the models to find four vulnerabilities in software used across EU institutions, all now fixed. A joint ENISA and CERT-EU report reviewed 27 software projects with frontier models; across the models tested, maintainers confirmed 95% of 88 assessed potential issues as genuine.
- CERT Polska used OpenAI models to help find six vulnerabilities in MikroTik RouterOS, with researchers verifying the findings. MikroTik has released fixes.
The update lands the same day California's attorney general sued OpenAI over its model testing practices, which we covered separately.
Blogs