WASHINGTON — The Pentagon has widened access to GenAI.mil, its secure artificial intelligence portal, to approximately 3 million military and civilian personnel. The expansion, announced under the Department of Defense’s AI Acceleration Strategy, now places three commercial AI models in the hands of the force for unclassified administrative, policy, and operational work. The move is a step to keep sensitive government data off consumer-grade systems.
The platform first launched in December 2025 with a single provider, and now carries Google Gemini, ChatGPT Mil, and Grok for Government as distinct options for users.
Three models, one secure door
Google Gemini remains the foundational model, cleared for Impact Level 5 (IL5) use since the platform’s December debut. ChatGPT Mil, built through OpenAI’s government program, is tailored for document-heavy work — planning, logistics, and policy development — with support for chat, file management, and custom GPTs. Grok for Government, supplied by Starshield AI (SpaceX/xAI) and also certified at IL5, leans into operational contexts: advanced reasoning, procurement market research, and supply chain management.
Data privacy is a core pillar. The platform is strictly exempt from standard consumer data harvesting and training models, and it is isolated from classified systems while certified to handle Controlled Unclassified Information (CUI).
The multi-model setup also sidesteps vendor lock-in, a point the Department of War made explicit: “The addition of Starshield AI’s Grok for Government to GenAI.mil provides another top-tier generative AI tool for warfighter advantage, eliminating vendor lock and promoting a vibrant American AI ecosystem.”
Scale and what comes next
The onboarding numbers are already substantial. According to the DoD, over 1.7 million unique users have signed onto the portal — a figure that underscores how quickly the force has taken to the tool since its December rollout. The expansion was reported by Federal News Network, with Michele Sandiford serving as producer for the Federal Drive.
For the personnel behind those logins, the practical effect is a faster lane for everyday tasks — drafting policy memos, analyzing procurement markets, or running logistics scenarios — without the risk of proprietary or sensitive information leaking into public AI systems. The deployment marks a major milestone in the Pentagon’s AI Acceleration Strategy and mirrors a broader federal push to fold generative AI into daily work under strict security controls.
Watch next: how the DoD balances the appetite for frontier models with the security reviews each new vendor must clear. With three providers now in the mix and millions of users onboard, the pressure will be on to keep the ecosystem both open and airtight.
