NATO has used up to 300 billion tokens a month of OpenAI cyber models, OpenAI says
OpenAI said NATO's core institutions have access to its advanced cybersecurity models through work with the alliance's communications agency under its Daybreak programme.

OpenAI said on 17 November 2026 that it works with the NATO Communications and Information Agency (NCIA) to give NATO’s core institutions access to its advanced cybersecurity models, in an announcement on its website. According to the company, NATO has used up to 300 billion raw input and output tokens a month as part of the work. It runs under Daybreak, the programme through which OpenAI gives trusted defenders access to its models for tasks such as reviewing software, validating vulnerabilities and testing fixes.
The announcement carried a statement from NCIA’s general manager, Dylan Browne, who said the models “accelerate vulnerability discovery”. He said that as adoption grows the agency would build an understanding of token consumption, usage patterns and the economics of AI models, to “optimise costs, evaluate competing solutions and make informed procurement decisions”.
OpenAI also said it supported the 2026 Crossed Swords exercise organised by the NATO Cooperative Cyber Defence Centre of Excellence. According to the company, the exercise included a new training stream in which participants built, deployed and evaluated AI agents for cyber operations, with an emphasis on human oversight.
Beyond NATO, OpenAI said it had given access to its cyber models to the EU cyber agency ENISA and to national cyber agencies in France, Germany and Poland. It pointed to a November joint report by ENISA and CERT-EU on a review of 27 software projects using frontier models, including OpenAI’s. Of the 220 potential issues the models found, software maintainers had assessed 88 by the time of publication, and 95 per cent of those were confirmed as genuine, the report says.
Source: OpenAI announcement