Grok and ChatGPT join Gemini in Pentagon’s enterprise genAI portal

The Pentagon’s suite of commercial generative AI products has expanded beyond Google’s Gemini offering, according to officials who told DefenseScoop Monday that insiders can now chat with approved versions of OpenAI’s ChatGPT and Starshield AI’s Grok frontier models when they log on to the GenAI.mil platform.

This long-anticipated integration is intended to enable U.S. personnel to apply the distinct technical strengths of each top-tier large language model provider to meet specialized military missions.

“Access to a diverse suite of AI capabilities from across the resilient American technology stack will give warfighters the tools they need to act with confidence and safeguard the nation against any threat,” said a source who asked to be referred to as a War Department official. 

(Officially changing the Defense Department’s name requires an act of Congress, but last year President Donald Trump signed an executive order rebranding DOD as the Department of War.)

Frontier AI encompasses the most advanced, large-scale, and capable foundational models that are pushing the boundaries of machine intelligence.

Within that realm is the emerging field of generative AI and associated large language models, which broadly produce convincing but not always correct text, software code, images, audio and other media from human prompts.

Commercial genAI products have maintained much hype since they were released to the general public in late 2022. However, experts have increasingly warned of severe known and unknown threats that the conversational tech poses to humanity, privacy, public trust, democratic processes and military operations.

In December 2025, the DOD unveiled plans for a swift and expansive release of the GenAI.mil platform to unleash advanced AI models to more than 3 million military members, civilian employees and contractors. Google Cloud’s Gemini for Government products were the first to go live in the one-stop shop. Officials at the time said that other models — including Anthropic’s Claude, OpenAI’s ChatGPT and xAI’s Grok — would soon follow. 

However, a dispute between the Pentagon and Anthropic derailed plans to add Claude to the platform. The company’s leadership insisted on stricter contractual guardrails that would prevent DOD from applying the firm’s AI technology to conduct mass surveillance of Americans or be used in connection to lethal autonomous weapons. The Pentagon, in turn, demanded unrestricted access to deploy any genAI products it purchased for what its leaders deem to be “lawful purposes.”

About six months ago, the department formally designated Anthropic a “supply-chain risk” and ordered personnel to purge the company’s AI products from all national security systems and assets. The Claude maker subsequently sued the government, and last week a federal district judge ruled that the Pentagon’s designation and actions against Anthropic were “illegal and baseless.”

“The Department of War will continue to build an architecture that prevents AI vendor lock and ensures long-term flexibility for the Joint Force,” the source who asked to be referred to as a War Department official said in response to questions from DefenseScoop Monday about the latest GenAI.mil announcements.

DOD uses an “Impact Level” classification system to ensure that cloud-based data hosting environments and services are safe and compliant with existing policies. 

The department’s press releases about Grok and ChatGPT joining Gemini on GenAI.mil indicate that each of the industry-leading frontier models have cleared Impact Level 5, or IL5, which covers the highest authorization granted to environments built to store and process non-public, sensitive unclassified data.

“We are moving at wartime speed: empowering the department to experiment, iterate, and deploy capabilities faster than our adversaries can adapt,” the War Department official said.

DOD’s initial announcement Monday revealed that OpenAI’s ChatGPT Mil “brings a familiar commercial experience” into the secure GenAI.mil environment that’s “tailored to warfighter needs.” In particular, the product will support document-heavy unclassified pursuits around planning, policy, logistics, and administration.

“The core experience centers on chat, files, projects, and custom GPTs, with additional features sequenced over time,” Pentagon officials wrote in the press release. 

A person familiar with the work at OpenAI said that DOD employees can now use ChatGPT Mil to generate reliable models for repeatable workflows — like filling out travel forms or responding to contracting opportunities.

“Personnel will soon be able to chat directly with GPT-5.6 Terra, OpenAI’s latest frontier model, bringing our most advanced model to research, analysis, drafting, and other day-to-day work,” the source told DefenseScoop, noting that the prior version 5.4 Terra is currently offered.

ChatGPT Mil also organizes relevant files and conversations that occur over months-long projects in one place, they added, as well as an “Offline Search capability” that provides detailed citations and can be refreshed to reflect new information as it emerges.

“This integration is grounded in our national security principles, pairing access to powerful tools with clear safeguards, meaningful human judgment, and accountability for high-stakes decisions,” Joe Larson, a former senior Pentagon official who now serves as OpenAI’s vice president and head of government, said in a statement.

Meanwhile, billionaire Elon Musk rebranded and merged his companies xAI and SpaceX into a combined corporate powerhouse earlier this year. 

Starshield AI’s Grok for Government is an xAI Grok frontier model that’s tailored for federal use. 

Per DOD’s announcement, Grok for Government will supply personnel with “advanced capabilities, including deep-thinking inference, adaptive reasoning modes (Auto, Fast, and Expert), customizable workspaces, persistent projects and reusable ‘playbooks’ that capture and scale institutional knowledge” via GenAI.mil. 

Spokespersons from xAI and SpaceX did not respond to requests for comment on Monday.