The Pentagon is giving its workforce two more AI assistants, adding custom versions of ChatGPT and Grok to its growing government AI platform.
The Department of War has added OpenAI’s ChatGPT Mil and Starshield AI’s Grok for Government to GenAI.mil, its centralized platform for accessing commercial generative AI tools in a more secure government environment.
Both models are accredited for Controlled Unclassified Information (CUI) at Impact Level 5, according to the Department of War. That allows personnel to use them for certain sensitive but unclassified work inside the government environment rather than relying on consumer versions of the services.
The expansion comes as GenAI.mil gains traction across the department. More than 1.7 million unique users have joined the platform out of a workforce of more than 3 million military and civilian personnel, the department said.
GenAI.mil initially launched with Google’s Gemini for Government and is now becoming a multi-model hub rather than a single-provider AI service.
ChatGPT targets document work as Grok adds reasoning modes
ChatGPT Mil is aimed largely at everyday work that consumes government employees’ time. According to the Pentagon, it supports chats, files, projects and custom GPTs, with uses including administration, logistics, planning and policy.
Grok for Government takes a more expansive approach. The Pentagon says it includes deep-thinking inference, three reasoning modes — Auto, Fast and Expert — customizable workspaces, persistent projects and reusable “playbooks” designed to preserve institutional knowledge.
The department said those capabilities could support work ranging from market research for acquisition professionals to supply-chain management for logistics personnel.
“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 of War said.
What eWeek found: Security approval does not decide which model to use
The Pentagon’s announcements show that GenAI.mil is becoming a governed multi-model environment rather than a single-provider AI service. ChatGPT Mil emphasizes document-heavy work, while Grok adds selectable reasoning modes, persistent projects and reusable playbooks.
Both tools have IL5 accreditation, but that approval addresses the environment and information they may handle — not which model is more accurate or appropriate for a particular task. The Pentagon must still establish rules for matching models to workloads and reviewing outputs, particularly when AI-generated information could influence consequential decisions.
For enterprises, the distinction is important: approving an AI platform to handle sensitive data is only one part of governance. Organizations also need workload-specific permissions, model-selection policies and human review requirements.
The Anthropic factor
The additions also land against the backdrop of the Pentagon’s increasingly contentious relationship with Anthropic.
The department designated Anthropic a national security supply-chain risk after negotiations over restrictions on the military use of Claude broke down. The dispute centered on potential applications including mass domestic surveillance and autonomous weapons.
A federal judge recently overturned that designation, ruling that the government's actions against Anthropic constituted unlawful retaliation, eWeek previously reported.
Before that ruling, The Hill reported that Pentagon Chief Technology Officer Emil Michael said the department expected to finish removing Anthropic’s platforms by the end of September. It is unclear whether the ruling will change that timeline, although Claude remains absent from the GenAI.mil lineup announced by the department.
Read more: OpenAI’s initial GenAI.mil agreement explains how ChatGPT became part of the Pentagon’s broader push to make commercial AI available across its workforce.