Core Event: Custom Models Are Now Live on GenAI.mil

The Pentagon has launched ChatGPT Mil from OpenAI and Grok for Government from xAI, adding both tools to GenAI.mil, the Department of Defense’s centralized and secure portal for AI tools. The portal is designed to give roughly 3 million civilian and military personnel access to commercial frontier AI models without routing sensitive government data through ordinary consumer channels.

GenAI.mil already offered Google Gemini when it first launched last year. According to the Defense Department, the portal has now onboarded more than 1.7 million unique users out of the department’s 3 million personnel.

Key facts:

• New tools: ChatGPT Mil and Grok for Government
• Platform: GenAI.mil, a centralized secure AI portal for the Department of Defense
• Existing model: Google Gemini was already available on the portal
• User base: roughly 3 million DoD civilian and military personnel
• Data policy: the military version is exempt from data collection that is difficult to avoid in consumer tech products
• Notably absent: Anthropic’s Claude is not part of the portal

Product Capabilities and Positioning

ChatGPT Mil comes from the OpenAI for Government program. The Defense Department says it will offer an experience familiar to commercial ChatGPT users, with a focus on:

• chat
• files
• projects
• custom GPTs

Its initial focus is document-heavy, routine unclassified work, including administrative tasks, logistics, planning, and policy-related work. The department said additional features will be added over time.

Grok for Government was described in broader and more operational terms. In the Defense Department’s press release, “Starshield AI’s Grok for Government will provide the warfighter with immediate productivity gains, stronger knowledge continuity and more secure and efficient collaboration.”

The department also said Grok would help the military execute missions faster and with greater precision across a range of operational contexts, from market research analysis for acquisition professionals to supply chain management for logisticians. The original report also notes that SpaceX’s Starshield AI is a secure satellite network using existing Starlink technology.

The Anthropic Gap and Industry Context

One notable contrast stands out: GenAI.mil has seen significant adoption, but Anthropic’s Claude remains absent.

According to the report, Anthropic was labeled a supply-chain risk by the Trump administration and is currently fighting that designation in court. The dispute followed Anthropic’s refusal to give the Pentagon unrestricted use of its AI tools; the company instead insisted on certain safety guardrails.

That absence highlights a broader tension in military AI procurement. The Pentagon wants the productivity and capability gains of frontier AI, but it also has to manage data protection, supplier risk, access terms, and safety constraints.

The Pentagon has also struck deals with other technology companies to strengthen its AI capabilities, including Amazon Web Services, Microsoft, Nvidia, and Reflection AI. The pattern suggests a multi-vendor strategy rather than reliance on a single model provider.

What It Means for Users and Observers

For DoD users, ChatGPT Mil appears best suited to routine unclassified workflows such as documents, administration, planning, logistics, and policy work. Grok for Government is being positioned more directly around warfighter productivity and mission execution. Both are being delivered through a secure portal meant to reduce the risks of using consumer AI services for government work.

For industry observers, the expansion shows how quickly frontier AI is moving into government and defense workflows. The competition will not be decided by model capability alone. Vendors will also need to meet government expectations around security, compliance, data separation, and acceptable-use boundaries.

Final Thought

By adding ChatGPT Mil and Grok for Government to GenAI.mil, the Pentagon is widening its internal use of generative AI. The real question is not only which model performs best, but which suppliers can balance capability, control, and security commitments at government scale.