The United States Department of Defense added OpenAI’s ChatGPT and xAI’s Grok to GenAI.mil on Monday 31 August, placing two more commercial systems alongside Google’s Gemini on the platform it uses for unclassified work. Both cleared the Impact Level 5 authorisation that covers controlled but unclassified information. The department said 1.7 million of its roughly 3 million personnel have used GenAI.mil since it launched with Gemini in December 2025.
The two additions are pitched at different jobs. The ChatGPT version is aimed at document-heavy tasks across planning, policy, logistics and administration. Grok for Government arrives with several reasoning modes, saved workspaces and reusable playbooks, with acquisition research and supply-chain work named as early uses. An official, not named in the announcement, said the aim was an architecture that prevents vendor lock and keeps long-term flexibility. No contract values were disclosed.
The name absent from the list is Anthropic. The Pentagon labelled the company a national security supply-chain risk this year after it declined to permit unrestricted military use of its Claude models, and a federal judge struck that label down the week before as retaliatory. So the case for spreading work across many American vendors arrived in the same fortnight the government lost in court over shutting one out.
What the tools are cleared to do is still not public. The announcement does not say whether operational planning, intelligence analysis or targeting support are in scope, or whether the carve-outs OpenAI agreed separately with the department, covering autonomous weapons and domestic surveillance, apply to this deployment. Nor is there a per-seat cost, or any breakdown of how much of that 1.7 million figure is repeat use rather than a single sign-in.
