Department of War launches OpenAI's ChatGPT Mil on GenAI.mil at IL5

The Department of War (formerly Department of Defense) has deployed ChatGPT Mil on its GenAI.mil platform, accredited to handle Controlled Unclassified Information at Impact Level 5 (IL5) — a high bar for sensitive government workloads. The deployment gives personnel access to OpenAI's GPT-5.6 Terra, and sits alongside xAI's Grok and Google's Gemini on the same platform, underscoring the government's model-agnostic procurement posture.

The launch is part of a rapid militarization of frontier models. It lands the same week the Pentagon rolled out Starshield AI with xAI's Grok for Government to 1.7 million users, and as Google's Gemini reached defense platforms. The convergence of three frontier labs onto Pentagon infrastructure within days signals that CUI-accredited AI has moved from pilot to procurement at scale.

The development raises governance questions around CUI accreditation, dual-use safety, and how autonomous agent capabilities interact with military security requirements — especially given the concurrent industry anxiety over agents autonomously executing attacks. For OpenAI, the deal deepens a commercial relationship with the U.S. government and validates the enterprise-grade safeguards it has been building into its model stack, even as its most advanced cyber features (Astra) remain gated to select partners. Watchers will focus on whether IL5 accreditation eventually extends to classified environments and how refusal-rate gating behaves in high-stakes operational contexts.