we_are_coded.by CODE · The world, decoded
БГ

OpenAI · Security

OpenAI and NATO: Daybreak enters the Alliance's cyber defence, and NATO has used up to 300 billion tokens a month

OpenAI is giving NATO's core institutions access to its cyber models through Daybreak; according to OpenAI, the Alliance has used up to 300 billion input and output tokens a month. ENISA found four vulnerabilities, CERT Polska six in MikroTik RouterOS.

Source
OpenAI
Published
The event
Checked on
Responsible editor
Cvetelin Ivanov
How we work
MethodCorrections

That is what NATO has spent working with OpenAI's models, according to the company itself: up to three hundred billion tokens a month, and the Alliance is inside the work.

Daybreak is the programme through which OpenAI gives trusted defenders access to its cyber models for authorised work: reviewing software, investigating suspicious activity, validating vulnerabilities, and developing and testing fixes. Today NATO's core institutions are in it too. Through the Alliance's Communications and Information Agency. Sasha Baker, Head of National Security Policy at OpenAI, puts it this way: the people defending our democracies should have the most advanced technology on their side.

Fact

The facts: on 17 November 2026 OpenAI announced it works with NATO's Communications and Information Agency (NCIA) to provide NATO's core institutions with access to advanced cybersecurity models. According to OpenAI, NATO has used up to 300 billion raw input and output tokens per month as part of this work. OpenAI also supported the 2026 Crossed Swords exercise, organised by the NATO Cooperative Cyber Defence Centre of Excellence (CCDCOE): it included a new training stream in which participants built, deployed and evaluated AI agents for cyber operations, using Daybreak in the technical environment. Beyond NATO, the company has provided access to its cyber models to defenders in Europe, including the EU's ENISA and national cyber agencies in France (ANSSI), Germany and Poland. According to OpenAI, ENISA used its models to identify four vulnerabilities in software used across EU institutions, with fixes published for all four. A November joint report from ENISA and CERT-EU describes a review of 27 software projects using frontier models including OpenAI models accessed through Daybreak, and tools including the Codex security plug-in; according to OpenAI, 95% of the 88 potential issues assessed by software maintainers were confirmed as genuine. CERT Polska used OpenAI models to help identify six vulnerabilities in MikroTik RouterOS and coordinated their disclosure with the vendor; MikroTik released fixes, and CERT Polska confirmed they prevented the attacks it had observed. The Ministry of Digital Transformation of Ukraine is using Daybreak to strengthen the cyber defence of civilian infrastructure.

There is no CVE number: this is how defenders use models, not a hole in a specific product.

Dylan Browne of NCIA calls it a generational leap in cyber security. He runs the agency that is using it.

CERT Polska did the hardest part the old way. People directed the work, tested the findings on real systems and verified their impact. The models accelerated the analysis. That is the difference between a tool and a replacement, and it shows in the report itself.

NATO is already inside the work. When the model finds something, the person still decides what to do.

The visual is generated code-art, no third-party images.