Pentagon's GenAI.mil Platform Now Adds ChatGPT and Grok Alongside Gemini
News Summary
The U.S. Department of War, formerly known as the Pentagon, has expanded its internal generative AI platform to include OpenAI's ChatGPT and xAI-affiliated Starshield AI's Grok, giving millions of military and civilian personnel a government-controlled way to use commercial-grade chatbots without sending sensitive data through public consumer apps. The rollout, announced on August 31, 2026 (Eastern Time), adds the two new tools to Google's Gemini, which had been the platform's sole large language model since its debut roughly nine months earlier.
What Is GenAI.mil
GenAI.mil is the Department of War's enterprise generative AI portal, designed to give the department's more than 3 million military members, civilian employees, and contractors access to leading commercial AI systems inside a secure, government-managed environment. Officials say the goal is to let personnel use familiar chatbot-style tools for everyday tasks such as drafting documents, summarizing reports, and organizing projects, while keeping the underlying data isolated from any public or commercial version of the models. According to reporting on the rollout, information processed through GenAI.mil does not get used to train or improve the companies' consumer-facing products.
The platform has already onboarded more than 1.7 million unique users out of the roughly 3 million personnel eligible to use it, according to figures released by the department at the time of the announcement.
New Tools Added: ChatGPT Mil and Grok for Government
The two newly added tools are branded specifically for government use. ChatGPT Mil, delivered through OpenAI's "OpenAI for Government" program, is positioned for document-heavy unclassified work such as planning, policy analysis, logistics coordination, and administrative tasks. It supports persistent chat threads, file uploads, long-running projects, custom GPTs, and offline search with citations, and officials say a newer underlying model is expected to become available to the department after the current version.
Grok for Government, offered through Starshield AI in connection with SpaceX, is described by the department as delivering "immediate productivity gains, stronger knowledge continuity and more secure and efficient collaboration" for military users. It includes deep-thinking inference modes, adaptive reasoning settings, customizable workspaces, persistent projects, and reusable "playbooks" meant to capture and scale institutional knowledge across teams.
Both tools join Gemini, which launched with the platform when it first went live and remains available for tasks such as search-oriented queries, giving personnel a choice of model depending on the task at hand.
Security Clearance and Data Handling
Each of the three large language models available on GenAI.mil has cleared Impact Level 5 (IL5) authorization, the U.S. government's security classification for environments approved to store and process non-public, sensitive unclassified information. IL5 is a high bar for cloud and AI systems handling government data, and clearing it for three separate frontier AI providers marks a notable expansion of accredited commercial AI options available to defense personnel. GenAI.mil is explicitly separate from classified AI systems the department is developing and testing for battlefield and operational use; it is scoped to unclassified administrative and analytical work.
The Anthropic Question
Notably absent from the platform is Anthropic's Claude, which had reportedly been under consideration for inclusion. According to multiple reports, talks between the department and Anthropic stalled over contractual terms: Anthropic sought guardrails limiting how its technology could be applied, while the department pushed for broader latitude to use the tools for its stated purposes. The dispute drew additional attention after the department had, months earlier, flagged Anthropic in connection with a supply-chain risk review — a designation a federal court reportedly found lacked adequate grounding shortly before the August 31 announcement. Anthropic has not been added to GenAI.mil as of this rollout, though officials have described the platform as open to additional providers over time as part of a broader push to avoid dependence on any single vendor.
Industry and Official Reaction
A War Department official framed the expansion as being about choice and resilience, saying access to a diverse suite of AI capabilities gives personnel the tools they need to work with confidence across a range of tasks. A representative from OpenAI's government affairs team described the arrangement as pairing powerful AI tools with clear safeguards, meaningful human oversight, and accountability for how the systems are used.
The department has also framed the multi-vendor approach as a way to support a broader domestic AI ecosystem and reduce reliance on any one company's technology, positioning GenAI.mil as an evolving, multi-model platform rather than a single-vendor deployment. Other technology partners associated with the broader GenAI.mil infrastructure include major cloud and computing providers that supply the underlying hosting and hardware for the platform.
What It Means for Government AI Adoption
The expansion illustrates how quickly large government agencies are moving to formalize internal access to commercial AI systems, treating them as productivity infrastructure similar to email or office software rather than experimental tools. For students and technologists watching AI adoption trends, the GenAI.mil rollout is a useful case study in how large organizations balance the convenience of frontier AI models against strict requirements around data security, auditability, and vendor accountability — challenges that extend well beyond the defense sector to any enterprise handling sensitive information at scale.