The Pentagon just got its own ChatGPT and Grok, minus Anthropic’s Claude

Roughly 1.7 million people already logged into the Pentagon’s AI portal before it had ChatGPT on it. That’s more than half of the Department of Defense’s 3 million civilian and military personnel, on a product most of them probably didn’t know existed a year ago.

Now the department has added ChatGPT Mil and Grok for Government to GenAI.mil, the centralized secure portal it launched last year. Both are custom builds of the consumer tools, tuned for what the department calls “warfighter needs.”

The absence is louder than the announcement

There’s no Claude here. Anthropic’s model is missing from a portal that now carries OpenAI’s and xAI’s, and that’s not an oversight.

Anthropic was labeled a supply-chain risk by the Trump administration after it refused to give the Pentagon unrestricted use of its AI tools and insisted on certain safety guardrails instead. The company is fighting that designation in court.

So the additions do double duty. They accelerate the department’s AI work, and they demonstrate that the Pentagon has other options after its dispute with the frontier lab.

What GenAI.mil is supposed to solve

The portal launched with Google Gemini. The pitch was straightforward: get Defense Department employees onto commercial frontier models without routing sensitive government data through ordinary consumer channels.

The official military version is exempt from the data collection that’s hard to escape in consumer tech products. That exemption is the entire product, arguably. Everything else is a model you can already use at home.

ChatGPT Mil is the boring one, and that’s fine

ChatGPT Mil comes out of OpenAI for Government, and it’ll feel familiar to anyone who’s used the commercial version. Chat, files, projects and custom GPTs, according to the department.

The stated use cases are document-heavy, routine unclassified work. Administrative tasks, logistics, planning, policy. More features are expected over time.

No claims about missions or precision. Just paperwork, which is honestly what most of 3 million people do all day.

Grok got the war-movie treatment

The Pentagon described xAI’s tool in noticeably different language.

“Starshield AI’s Grok for Government will provide the warfighter with immediate productivity gains, stronger knowledge continuity and more secure and efficient collaboration,” the department’s press release said. Starshield AI is SpaceX’s secure satellite network, built on the company’s existing Starlink tech.

The department added that Grok would let its military “to execute missions faster and with greater precision across numerous operational contexts, ranging from market research analysis for acquisition professionals to supply-chain management for logisticians.”

Read that list again. The operational contexts are market research and supply-chain management. The framing is battlefield; the examples are procurement.

The vendor list keeps growing

The Pentagon has signed deals with Amazon Web Services, Microsoft, Nvidia and Reflection AI as part of the same push.

If you work at the department, the practical question isn’t which model is smartest. It’s whether the secure version of ChatGPT you get at work does the thing you’d otherwise paste into the consumer app on your phone, because that’s the behavior GenAI.mil exists to stop.