Key Takeaways
The Defense Department added OpenAI's ChatGPT Mil and Starshield AI's Grok for Government to its GenAI.mil platform on August 31, joining Google's Gemini.
GenAI.mil now serves a workforce of more than 3 million, with about 1.7 million users already onboarded across the department.
Anthropic's Claude is the only major frontier model still absent, months after the Pentagon labeled the company a supply-chain risk and a federal judge called that move illegal.
The Department of Defense added OpenAI's ChatGPT and Starshield AI's Grok to its secure generative AI platform, GenAI.mil, on Monday, August 31, DefenseScoop reported. The two frontier models join Google's Gemini, which was the first commercial model to go live on the platform earlier this year.
GenAI.mil is cleared to handle sensitive but unclassified data at Impact Level 5, the highest authorization for information of that kind. The Pentagon built it for a workforce of more than 3 million service members, civilian employees and contractors, and says about 1.7 million people have already signed on. OpenAI's version, branded ChatGPT Mil, is aimed at document-heavy work like planning, logistics and administration, and Grok for Government adds reasoning modes and reusable playbooks.
One name is missing. Anthropic's Claude is the only major frontier model still absent from GenAI.mil. The company pushed for contract terms that would keep its models out of mass surveillance of Americans and lethal autonomous weapons, and the Pentagon wanted unrestricted access for what it calls lawful purposes. About six months ago the department designated Anthropic a supply-chain risk and ordered its products purged from national security systems.
Anthropic sued, and last week a federal district judge ruled the designation was "illegal and baseless." A defense official told DefenseScoop the department is building an architecture that prevents AI vendor lock and keeps long-term flexibility for the force.
The pattern here is hard to ignore. The government is standardizing on the models it can direct without limits, and steering around the one vendor that asked for guardrails. It is the same tension the newsroom has followed as Washington draws lines around AI, from the FDA weighing how to test generative models the way it tests a doctor to the national-security scrutiny of which AI systems the government will and will not touch. Worth watching whether the contracts or the courts settle where those lines land.
People Also Ask
What is GenAI.mil?
GenAI.mil is the Defense Department's secure generative AI platform for unclassified work. It now hosts Google's Gemini, OpenAI's ChatGPT Mil and Starshield AI's Grok for Government, all cleared at Impact Level 5.
Why is Anthropic's Claude not on GenAI.mil?
A contract dispute. Anthropic wanted terms barring mass surveillance and autonomous weapons use, the Pentagon wanted unrestricted access, and the department later designated Anthropic a supply-chain risk.
How many people use GenAI.mil?
The Pentagon says about 1.7 million users have onboarded out of a workforce of more than 3 million service members, civilian employees and contractors.
What did the court rule about the Pentagon's Anthropic ban?
A federal district judge ruled last week that the department's supply-chain-risk designation and purge order against Anthropic were illegal and baseless.
Sources
DefenseScoop, U.S. Department of War, FedScoop, Defense One.