Pentagon's GenAI.mil Adds ChatGPT Mil and Grok for Government for 3 Million Users (2026)

The Pentagon's GenAI.mil portal adds ChatGPT Mil and Grok for Government for its three million military users — and why Anthropic's Claude is absent from the platform.

Pentagon's GenAI.mil Adds ChatGPT Mil and Grok for Government for 3 Million Users (2026)
Table of contents
Last updated: September 2026

Three million people. That is how many U.S. service members, civilian employees and contractors can now reach military-grade versions of ChatGPT and Grok through a single secure gateway, after a Department of Defense announcement on Monday evening, August 31, 2026.

On Monday, August 31, 2026, the U.S. Department of Defense announced that ChatGPT Mil from OpenAI and Grok for Government from Starshield AI joined Google's Gemini for Government on the GenAI.mil portal. The platform, which has already onboarded more than 1.7 million unique users, is the Pentagon's central gateway to commercial AI models.

The Pentagon, headquarters of the U.S. Department of Defense

Source: Wikimedia Commons

What GenAI.mil is

GenAI.mil is the Pentagon's central secure portal for commercial artificial intelligence. Plans for the gateway were announced in December 2025, with a clear purpose: give more than 3 million military service members, civilian employees and contractors access to advanced commercial models without sensitive government data ever passing through ordinary consumer channels.

Google's Gemini for Government was the first model to join the portal. The military versions of these tools are exempt from the data collection practices that come with consumer products, which is the core reason a separate gateway exists at all. A defense employee drafting logistics notes should not have those notes feeding a public chatbot's training pipeline.

The adoption numbers released so far are striking. Beyond the 3 million people eligible, the department says the platform has onboarded more than 1.7 million unique users to date — a figure that makes this one of the largest organized deployments of commercial AI models anywhere in the public sector.

Aerial view of the Pentagon

Source: Wikimedia Commons

What was added on Monday

Two models joined the portal on August 31, 2026, according to the official announcement reported by DefenseScoop and TechCrunch, the latter publishing its story at 20:13 UTC on Monday evening.

ChatGPT Mil came out of the OpenAI for Government program. The department describes it as a familiar experience for anyone who uses the commercial ChatGPT: conversations, files, projects and custom GPTs are all there. It is intended for routine unclassified work — administrative tasks, logistics, planning and policy — with additional features to be added over time.

Grok for Government arrives through Starshield AI, the SpaceX unit that operates a secure satellite network built on Starlink technology. Grok itself is the AI model built by xAI, whose background we have covered in detail. The official statement sets a high bar: "Grok for Government from Starshield AI will deliver immediate productivity gains, stronger knowledge continuity, and more secure and efficient collaboration for the warfighter."

A department official put the wider strategy in plain terms, telling DefenseScoop: "Access to a diverse set of AI capabilities from across the resilient American tech stack will give warfighters the tools they need to act with confidence and protect the nation from any threat."

One naming note for careful readers. Officials have used "Department of War" after an executive order renamed it last year, but the formal change requires a decision by Congress — which is why the institution is still officially the Department of Defense.

Quick comparison

Model on GenAI.mil Provider What was announced Status
Gemini for Government Google First model to join the secure portal Live since the portal's launch
ChatGPT Mil OpenAI Chats, files, projects and custom GPTs for routine unclassified work: administration, logistics, planning, policy Added August 31, 2026
Grok for Government Starshield AI, a SpaceX unit Productivity gains, knowledge continuity and secure collaboration for the warfighter Added August 31, 2026
Claude Anthropic Not part of Monday's addition Absent amid a legal dispute

The Claude question

The missing name on that list is Anthropic's Claude, and the backstory explains a great deal about how governments and AI companies negotiate. Anthropic refused unrestricted use of its tools and insisted on safety guarantees. The administration then classified the company as a supply chain risk, and Anthropic took the classification to court.

On August 28, 2026 — three days before Monday's launch — federal judge Rita F. Lin of the Northern District of California granted Anthropic summary judgment, finding the classification unlawful. We covered that ruling in full when it landed. Claude was still nowhere in Monday's announcement, which tells you the repair job between the two sides is not finished.

What this means for you

Why should a reader in Riyadh, Lagos, Jakarta or Warsaw care about a U.S. military portal? Because Monday's announcement is a signal about where institutional money and trust are flowing. When a government of this size makes commercial AI models standard infrastructure for 3 million people, it confirms that generative AI has moved from experiment to procurement — and that the investment behind it, from data centers to satellite networks, is budgeted for the long term.

For professionals outside the United States, two practical points stand out. First, nothing changes for ordinary users: the military versions are separate products, and your personal ChatGPT or Grok account does not interact with them. Second, there is a business lesson in the Claude story — even the largest government buyer on earth faces conditions when it signs AI suppliers, and safety terms can derail a mega-contract. If you advise organizations on AI adoption, vendor terms deserve the same scrutiny as model quality.

An event inside the Pentagon

Source: Wikimedia Commons

What we still do not know

The announcement leaves real gaps. The financial terms are undisclosed — no contract values or pricing have been published. The precise contents of the security agreements between the department and the three companies are not public. And while officials say more features will come to ChatGPT Mil over time, no dates were offered for any of them.

Frequently asked questions

What is GenAI.mil, the Pentagon's AI portal?

GenAI.mil is the U.S. Department of Defense's central secure gateway to commercial AI models, first announced in December 2025. It exists so that more than 3 million service members, civilian employees and contractors can use advanced models without sensitive government data passing through ordinary consumer channels.

What is ChatGPT Mil and how does it differ from regular ChatGPT?

ChatGPT Mil is the government version that grew out of the OpenAI for Government program. It feels like the commercial product — chats, files, projects and custom GPTs — but it is intended for routine unclassified work such as administration, logistics, planning and policy. More features are expected over time.

What is Grok for Government by Starshield AI?

Grok for Government is the military version of Grok, delivered through Starshield AI, a SpaceX unit operating a secure satellite network built on Starlink technology. The Pentagon said it will deliver immediate productivity gains, stronger knowledge continuity, and more secure and efficient collaboration for the warfighter.

Why is Anthropic's Claude missing from GenAI.mil?

Anthropic refused unrestricted use of its tools and insisted on safety guarantees. The administration then classified the company as a supply chain risk, which Anthropic challenged in court. On August 28, 2026, Judge Rita F. Lin granted summary judgment finding the classification unlawful. Claude was not part of Monday's addition.

How many users does GenAI.mil have?

More than 3 million people are eligible: military service members, civilian employees and contractors. According to the department, the platform has onboarded more than 1.7 million unique users so far, and Monday's announcement extends the set of models available to all of them.

What to watch next

Three things decide whether this story grows or stalls. Watch whether the promised feature additions to ChatGPT Mil actually arrive, and how fast. Watch whether Anthropic and the department repair their relationship after the August 28 court ruling — a fourth major model on the portal would be a significant shift. And watch for any disclosure of contract values, which would reveal how much of this is substance versus symbolism.

For now, the direction is unmistakable: the world's largest institutions are standardizing on commercial AI, and the companies building it just gained their biggest anchor customer yet. Keep following Truescho for clear, jargon-free reporting on the AI stories that move markets, governments, and your own work.

Sources