OpenAI launches ChatGPT gov

OpenAI introduced ChatGPT gov, a version of ChatGPT built for US federal agencies to use with sensitive but unclassified data. Rather than running on OpenAI’s own infrastructure, agencies would self-host it within Microsoft’s Azure commercial cloud or Azure Government cloud, with an administrative console for IT teams to manage access, aiming to meet security frameworks including FedRAMP High, IL5, CJIS and ITAR. Formal FedRAMP authorisation was still pending at launch, alongside separate accreditation work for ChatGPT Enterprise.

OpenAI said the launch built on existing federal use: it reported more than 90,000 users across roughly 3,500 federal, state and local agencies had already exchanged over 18 million messages with ChatGPT since 2024, with agencies including NASA, the Treasury Department and USAID among adopters of ChatGPT Enterprise. Coverage at the time described it as OpenAI’s biggest government-facing launch since its enterprise product, and noted the company was ahead of rivals — Anthropic was pursuing comparable federal access separately, through partnerships with Palantir and AWS.

The launch set a pattern OpenAI and its competitors repeated through 2025: government-tailored deployments followed months later by OpenAI’s $1-per-agency ChatGPT Enterprise offer to the entire federal workforce and Anthropic’s matching deal extending Claude to all three branches of government, as both labs competed for a foothold in public-sector procurement.