Anthropic trades its $1 government deal for 40% off Claude, from Washington to city hall

Anthropic’s $1 deal for US federal agencies runs out on October 31, and the company is replacing it with a discount open to every level of government. Under a new agreement with the General Services Administration (GSA), announced on October 27, government buyers get 40% off Claude Enterprise, Claude for Government and the Claude API, with no minimum spend.

The GSA’s own release describes it as a “40% discount on token-based usage across Claude models available to the government”, including FedRAMP-authorised environments for sensitive unclassified work. Agencies pay only for what they use.

Who can buy, and how

The GSA says the offer covers all three branches of the federal government (executive, legislative and judicial) plus state, local and tribal governments; Anthropic’s post adds territorial governments too. Anthropic says this is the first time the discount applies on every product surface: agencies buying directly from Anthropic get it now, and those buying through cloud providers will get it “in the coming months”. The GSA lists direct sales, resellers and major cloud platforms as ordering routes, and agencies can order through its Multiple Award Schedule today.

Anthropic estimates the discounts will save the government $1 billion, a company projection rather than a measured figure. Neither release says how long the new offer lasts. Anthropic is also starting an implementation programme to help agencies put Claude to work, borrowing what it does for business customers.

From $1 a seat to paying by the token

The deal it replaces was very different. In August 2025 the GSA struck a OneGov deal giving federal agencies Claude for Enterprise and Claude for Government for a nominal $1, for up to a year. That offer, later extended, ends on October 31.

OpenAI made the same switch first. Its new OneGov agreement, announced in September, gives governments 50% off token-based ChatGPT usage for 27 months from October 1, a bigger discount than Anthropic’s. When that deal was announced, the GSA official in charge of federal buying explained the shift:

We launched OneGov AI deals because agencies wanted the ability to try new technologies before committing to a lengthy acquisition. As agencies increasingly integrate AI into their regular operations, providing consumption-based access is the next logical step.

Laura Stanton, acting commissioner of GSA’s Federal Acquisition Service, in September

The GSA says OneGov has saved more than $1.8 billion since it began, about $1.6 billion of it through what its release calls “super intelligence agreements”, the wording Trump’s September executive order told agencies to use in place of “AI”. Unlike the OpenAI announcement, which quoted GSA Administrator Edward C. Forst and Sam Altman, the Anthropic release quotes no one.

What Anthropic says government does with Claude

Anthropic says about 500 US federal, state and local government organisations now use Claude, that weekly active users of Claude for Government are up about eightfold since February, and that usage across federal civilian agencies has grown about 20-fold in a year. None of those figures has been independently confirmed.

Its examples are unnamed apart from one. Utah gave Claude Code to 34 developers across 10 state agencies in a four-week pilot, and Anthropic says 71% reported saving 20 hours or more. It also says a federal security team used Claude Mythos to find and patch, in about two days, a flaw that would have let attackers take over a critical production system, one a human penetration test had missed the month before. Claude for Government became generally available at the FedRAMP High level on September 30.

Despite the fight with the Pentagon

The deal widens Anthropic’s government business while part of the government is still trying to shut it out. In February President Trump ordered agencies to stop using Anthropic’s technology, and the Pentagon branded it a supply chain risk; an appeals court let that blacklist stand in September and the Pentagon has stopped using Claude. Civilian buyers have kept going: last week the Department of Health and Human Services moved to award Anthropic a no-bid contract, saying only Claude would do.

The agreement also lands the day after Anthropic filed its S-1 publicly for a Nasdaq listing, where investors will be weighing how much revenue the US government can bring in.

Why it matters

The free-trial era of government AI is ending: agencies now pay for what they use, so Claude has to earn its place on budgets rather than ride a $1 deal. Opening the discount to states, cities and tribal governments gives Anthropic a much bigger public-sector market, even as the Pentagon keeps it out.

Sources: Anthropic, GSA (October 27, 2026), GSA (OpenAI agreement), GSA (August 2025).