OpenAI Daybreak Reaches NATO and EU Cyber Agencies, With Up to 300 Billion Tokens a Month

OpenAI said on 17 November 2026 that NATO's Communications and Information Agency uses its Daybreak cyber models at up to 300 billion tokens a month, and that ENISA, ANSSI and CERT Polska used them to find software vulnerabilities.

Christopher Ort
By Christopher Ort · 2 min read
Share
An empty security operations room with monitors and a headset on a desk.
Image: AI-generated, i10X
On this page
  1. Quick Take
  2. In plain English
  3. What happened
  4. Background
  5. How it works
  6. Why it matters
  7. What's still open
  8. Sources
  9. i10x Perspective

OpenAI has detailed how its Daybreak cyber defense program is used by NATO and by European national cyber agencies. In a post on 17 November 2026, the company said it works with NATO’s Communications and Information Agency (NCIA), which has used up to 300 billion raw input and output tokens per month, and has given cyber models to defenders at ENISA and agencies in France, Germany and Poland.

i10X Quick Take

Quick Take

  1. NATO: OpenAI gives NATO’s core institutions access to advanced cybersecurity models through the NCIA. Usage reached up to 300 billion tokens per month.
  2. Exercise: Daybreak was used in the 2026 Crossed Swords exercise run by the NATO Cooperative Cyber Defence Centre of Excellence, with participants building and testing AI agents for cyber operations.
  3. Other agencies: ENISA, France’s ANSSI, Germany and Poland have access, and Ukraine’s Ministry of Digital Transformation uses Daybreak for civilian infrastructure.
  4. Results: ENISA found four vulnerabilities in software used across EU institutions; CERT Polska found six in MikroTik RouterOS.
  5. Joint report: in a review of 27 software projects, 95% of 88 potential issues assessed by maintainers were confirmed as genuine.

In plain English

Security teams at governments use AI models to read code, spot weaknesses and test fixes faster. OpenAI says its Daybreak program gives vetted defenders these tools, and the post lists who is already using them and what they found. Humans direct the work and verify each finding.

What happened

OpenAI said the NATO work “helps protect Allied networks” and that Crossed Swords included a new training stream on AI agents for cyber operations, with an emphasis on maintaining human oversight. Dr Dylan Browne, NCIA General Manager, said: “As we scale adoption, we will build a deeper understanding of token consumption, usage patterns and the underlying economics of AI models.”

Background

Daybreak is OpenAI’s program that gives verified defenders access to its more capable cyber models. Earlier this month the company introduced GPT-6 Cyber for Daybreak Red customers and said it wanted to expand Daybreak to more organizations. See our earlier coverage of the Daybreak red team and defense livestream.

How it works

Defenders use the models for authorized work such as reviewing software, investigating suspicious activity, validating vulnerabilities and developing and testing fixes. In the CERT Polska case, researchers directed the work, tested findings on real systems and verified their impact, and MikroTik released fixes that CERT Polska confirmed prevented attacks it had observed.

Why it matters

For public-sector security teams, the post gives concrete usage numbers and results that can inform procurement decisions.

For OpenAI, government cyber work is a growing business line alongside its enterprise products.

For software maintainers, AI-found bugs mean more disclosures to triage, and the 95% confirmation rate suggests the reports are mostly real.

What’s still open

  • What NATO and the agencies pay for access, which OpenAI did not disclose.
  • How performance compares with other vendors’ cyber models, as NATO says it will evaluate competing solutions.
  • How oversight rules for AI agents in cyber operations will be set.

Sources

  • OpenAI, “Strengthening NATO and allied cyber defenses with AI,” 17 November 2026

i10x Perspective

The post is vendor-written, so the numbers deserve cross-checking, but the NATO quote about weighing costs and competing solutions shows buyers are already comparing vendors. See more in our Partnerships & Deals section and on the OpenAI company page.

Christopher Ort

Christopher Ort

Christopher Ort handles SEO, AI agent workflows and PR & partnerships at i10x.ai, the all-in-one AI workspace built by i10X Global Pte. Ltd. in Singapore. He is a graduate of Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). He is the author of “The Style Penalty”, a study published on SSRN that analyses 1,576 evaluation decisions by four frontier AI models to show how AI-written CVs fare in automated hiring screens; the research has been covered by heise online and t3n.

Timeline

What happened before

  1. Release

    GPT-6.1 Terra

    1.05M context · $1 / $5 per 1M tokens

    openai.com (source, opens in a new tab)
  2. Acquisition

    OpenAI acquires Orchestra Bio

    Acquires Orchestra Bio

    openai.com (source, opens in a new tab)
  3. Release

    GPT-6.1 Sol

    1.05M context · $2 / $10 per 1M tokens

    developers.openai.com (source, opens in a new tab)
  4. Release

    GPT-6 Sol & Luna

    1.05M context · $2 / $10 per 1M tokens

    Two GPT-6 reasoning models. Price shown is GPT-6 Sol (≤272K prompt); GPT-6 Luna costs $0.10 / $0.50.

    developers.openai.com (source, opens in a new tab)
  5. Release

    GPT-6 Astra

    1.05M context · $10 / $50 per 1M tokens

    OpenAI's most capable model for reasoning, coding, computer use and research.

    openai.com (source, opens in a new tab)
Full OpenAI timeline ›

Editor's picks

  1. Microsoft: device code phishing can hijack AI coding agents like GitHub Copilot CLI
  2. Trinity College Philosophy Department Bans AI Use in All Courses, for Any Purpose
  3. Hour of AI 2026: CodeAI, Formerly Code.org, Launches 250 Free AI Activities for Students
  4. Qatar Launches a National AI Tutor for Schools and an AI Academic Adviser With Scale AI
  5. Novo Rebuilds Small Business Banking Around AI as Its Customers’ AI Tool Spending Jumps 493%