Pentagon adds military-grade ChatGPT and Grok to secure AI platform

The Pentagon announced that its GenAI.mil platform will host military-grade versions of ChatGPT and Grok, enabling service members and civilian staff to process Controlled Unclassified Information at Impact Level 5. The platform, which already supports a government-specific Google Gemini, now serves 1.7 million users out of the department’s three-million workforce. OpenAI’s ChatGPT Mil will keep all data within the secure environment and will not use it for model training, while Starshield AI’s Grok for Government offers multiple reasoning modes for tasks such as market-research analysis and supply-chain management.

These additions stem from contracts of up to $200 million each awarded to OpenAI and SpaceXAI, alongside separate deals with Google and Anthropic. Secretary Pete Hegseth has been urging personnel to embed AI tools into daily workflows, aligning with the White House AI Action Plan’s call for aggressive AI adoption in the armed forces. Anthropic’s Claude is absent after the department labeled the company a supply-chain risk, a decision later ruled unlawful by a judge.