American service members will now have access to military versions of ChatGPT and Grok as the Pentagon moves to incorporate more artificial intelligence into its daily workflows.

The Department of War said in two announcements on Monday that the services will be added to its custom AI platform, GenAI.mil, which launched in December and initially offered access only to a specialized military version of Google’s Gemini. The department said GenAI.mil already has 1.7 million users out of the Department of War’s 3 million-person workforce, which includes both military members and civilians. That puts the platform at the center of an unusually large internal rollout, with the latest additions extending access to tools already built for government use rather than general consumer chatbots.

Monday’s announcement follows the department’s awards of defense contracts worth up to $200 million each to OpenAI and Elon Musk’s SpaceXAI, which operates Grok, to provide AI tools to the department. At the same time, contracts were also awarded to Google and Anthropic.

The military version of ChatGPT, called ChatGPT Mil, is part of OpenAI for Government, an initiative launched last summer to bring the company’s AI products to government workers.

“ChatGPT on GenAI.mil will make unclassified work throughout the Department more efficient, freeing up time and resources so DoW personnel can focus on the mission,” Joe Larson, vice president and head of government at OpenAI, said in a statement to Fortune.

“ChatGPT Mil brings a familiar commercial experience into the Department’s secure environment, tailored to warfighter needs,” the Pentagon said in a press release.

The specialized version of ChatGPT, which helped launch the AI race when it debuted in 2022, will keep data within its environment and will not use that data to train its models. According to the Pentagon announcement, service members can use both ChatGPT Mil and Grok for Controlled Unclassified Information (CUI) at Impact Level 5 (IL5), which refers to sensitive U.S. government information that is not classified. That security layer is important because much of the department’s day-to-day work happens outside classified settings, where routine approvals, records, logistics and procurement still require handling sensitive information.

The vast majority of military work is unclassified and centers on document-heavy routine tasks in human resources, logistics and contracting, much of which could be supported by ChatGPT’s ability to process large amounts of information, according to a person familiar with uses at OpenAI.

For Grok, the military version is described as Starshield AI’s Grok for Government. The Pentagon said the tool will give personnel access to deep thinking inference and multiple reasoning modes, including auto, fast and expert. It will also allow “the Joint Force 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,” according to the announcement.

The release of ChatGPT and Grok for the military comes as War Secretary Pete Hegseth has pushed for AI to become part of the Pentagon’s daily workflows, in line with the White House’s AI Action Plan, which says, “The United States must aggressively adopt AI within its Armed Forces if it is to maintain its global military preeminence.”

A War Department official said the department is adopting an “AI-first” mentality in the military that will transform how it fights, thinks and organizes to guarantee American dominance and security.

Even so, reporting by DefenseScoop found in December that the initial rollout of GenAI.mil, which at the time included only a military version of Google’s Gemini, was rushed, with some service members and workers unaware of the initiative at launch. The expansion to ChatGPT and Grok suggests the platform is still being built out in stages, even as the department broadens access to more vendors inside the same secure environment.

‘I want YOU to use AI’

On Dec. 9, the day of the rollout, workers saw a pop-up on their computers with a stylized GenAI.mil logo and the words “I want YOU to use AI,” along with a link to a government website, DefenseScoop reported. The style and suddenness of the launch led at least one unnamed senior Army official and people in their office to wonder whether the pop-up might be the result of a malicious actor hacking their computers.

According to DefenseScoop, the concern was not cleared up until DoW personnel saw a memo circulated by Hegseth, along with social media videos in which he said, “The future of American warfare is here, and it’s spelled AI.”

Hegseth’s December memo said, “I expect every member of the Department to log in, learn it, and incorporate it into your workflows immediately. AI should be in your battle rhythm every single day.”

Notably absent from Monday’s expansion of GenAI.mil is Anthropic’s Claude. The AI company led by Dario Amodei had a falling out with the Department of War that became public in January, when Hegseth, while announcing that the Pentagon would add Grok to its list of generative AI providers, criticized AI models that “won’t allow you to fight wars.” That comment was later revealed to be aimed at Anthropic over the company’s push for the Pentagon to agree not to use its technology for autonomous weapons or surveillance during contract negotiations.

In February, President Donald Trump directed federal agencies to stop using Anthropic’s technology in a process meant to conclude no later than the end of this month. Hegseth said at the time he would designate the company a “supply-chain risk.” Anthropic sued a month later.

Last week, a judge ruled that the Department of War acted unlawfully when it designated Anthropic a supply-chain risk.

This story was originally featured on Fortune.com