Anthropic is taking direct aim at OpenAI in the battle for federal AI adoption. Just a week after OpenAI announced ChatGPT Enterprise would be available to the entire U.S. federal executive branch for $1 per year per agency, Anthropic revealed Tuesday it will match the $1 price and expand it to all three branches of government, including legislative and judiciary agencies.
The offer, valid for one year, includes both Claude for Enterprise and Claude for Government, the latter certified for FedRAMP High workloads, the highest security baseline for handling unclassified sensitive government data. Anthropic says this ensures agencies can use Claude for everything from scientific research to constituent services without compromising security.
“By combining broad accessibility with uncompromising security standards, we’re helping ensure AI serves the public interest,” Anthropic said in a statement.
The move follows the addition of Anthropic, OpenAI, and Google DeepMind to the General Services Administration’s list of approved AI vendors for civilian federal agencies. Anthropic will also provide technical support to help agencies integrate AI into their workflows.
Related: OpenAI Drops GPT-5 With Pricing That Could Shake the AI Market
Backed by Department of Defense funding of up to $200 million for national security AI applications, Anthropic is already in use at Lawrence Livermore National Laboratory to accelerate research and at the District of Columbia Department of Health to deliver multilingual health services.
Anthropic’s multicloud partnerships with AWS, Google Cloud, and Palantir could offer an edge over OpenAI’s FedRAMP High option, which is currently tied to Azure Government Cloud. While Azure is widely adopted, some agencies may prefer the flexibility, data sovereignty, and infrastructure control that multicloud access provides.
With both companies undercutting traditional enterprise AI pricing, the race to secure government AI contracts is rapidly intensifying, and Anthropic is making sure it’s playing on every field.
