LATEST
AI Tools

Anthropic Ends $1 Claude for Government Today; 40% Off Next

Updated October 31, 2026
10 min read
Editor's Note: This article was human-reviewed for accuracy
US Capitol dome with an American flag in Washington, DC

Quick Answer & Key Takeaways

  • • Anthropic's $1 Claude pricing for US federal agencies expires on October 31, 2026.
  • • From November 1, federal, state, local, tribal and territorial agencies get a standing 40% discount on Claude Enterprise, Claude for Government and the Claude API, with no minimum spend.
  • • Anthropic estimates the deal will save governments $1 billion; buying is live now through GSA's Multiple Award Schedule.

Today is the last day of $1 Claude for the US government. Since August 2025, federal agencies have been able to put Claude in front of their staff for a nominal fee. That offer expires tonight, October 31, 2026, and from tomorrow the meter starts running.

Four days ago, Anthropic and the US General Services Administration announced what comes next: a standing 40% discount on token-based Claude usage for federal, state, local, tribal and territorial governments, with no minimum spend. It is the second government-wide AI pricing deal in two months, following OpenAI's 50% offer in September, and it moves the biggest public-sector AI buyer from flat-fee pilots to metered billing.

Here is what changes on November 1, what the 40% discount is worth in dollars, pounds and Canadian dollars, and the questions buyers should get answered in writing before the first metered invoice arrives.

What expires today: the $1 era

On August 12, 2025, the GSA announced its original OneGov agreement with Anthropic: Claude for Enterprise and Claude for Government for federal agencies at a nominal fee. The original GSA release described it as $1 per agency; GSA's OneGov listing later described the offer as $1 per user for executive-branch agencies. The two descriptions have circulated side by side ever since, and neither agency has clarified the discrepancy publicly.

The deal covered all three branches of the federal government and was supposed to run for a year, ending September 30, 2026. In September, GSA extended it by a month, moving the end date to October 31, accordng to a listing on GSA's OneGov website reported by Washington Technology. That extension runs out tonight.

The $1 pilot covered Claude for Enterprise and Claude for Government. It did not cover the Claude API, which matters because the replacement deal does.

What replaces it: 40% off, no minimums

On October 27, 2026, the GSA and Anthropic announced an expanded OneGov agreement. The headline terms, from the GSA's release and Anthropic's announcement:

  • A 40% discount on token-based usage across all Claude models available to the government, including FedRAMP-authorized environments for sensitive unclassified work.
  • No minimum orders and no spend commitment. Agencies pay only for the usage they consume.
  • Coverage across three product surfaces for the first time: Claude Enterprise, Claude for Government, and the Claude API.
  • Eligibility beyond the federal government. Federal buyers in the executive, legislative and judicial branches qualify, and for the first time so do state, local and tribal governments. Anthropic's announcement adds territorial governments to the list.
  • Buying is open now through GSA's Multiple Award Schedule, directly from Anthropic and through resellers. Cloud-provider purchases, through Amazon Bedrock and Google Cloud's Vertex AI, follow in the coming months. Anthropic says MAS ordering opened the day of the announcement.

GSA lists the new consumption offer as running through January 31, 2029, a little over two years. Anthropic estimates the discounts will save governments about $1 billion, a company estimate that depends entirely on how much gets bought.

Anthropic paired the announcement with its own usage snapshot: roughly 500 US federal, state and local government organizations now use Claude, weekly active users of Claude for Government are up about eightfold since February, and usage across federal civilian agencies has grown about twentyfold in a year. Those are company-reported figures with no independent confirmation, but they explain why the GSA moved from a pilot to a standing discount.

What 40% off means in dollars

The government-specific list prices are not published, so the honest way to read the discount is to apply it to Anthropic's published commercial API rates and treat the result as directional. The table below uses the commercial rate card as published on Anthropic's official API pricing page, with sterling and Canadian-dollar equivalents at October 31, 2026 market rates of £0.747 and C$1.397 per US dollar (verified against live market data at publication).

ModelCommercial list (per 1M tokens, input / output)At 40% off (USD)Approx. GBPApprox. CAD
Claude Opus 5.5$4.00 / $20.00$2.40 / $12.00£1.79 / £8.96C$3.35 / C$16.76
Claude Sonnet 5.5$2.00 / $10.00$1.20 / $6.00£0.90 / £4.48C$1.68 / C$8.38
Claude Haiku 5.5 (prompts up to 100K tokens)$0.10 / $0.50$0.06 / $0.30£0.04 / £0.22C$0.08 / C$0.42
Claude Fable 5.1$10.00 / $50.00$6.00 / $30.00£4.48 / £22.41C$8.38 / C$41.91

Two pieces of fine print travel with this table. First, Haiku 5.5's cheap tier only applies to prompts up to 100,000 tokens; longer prompts bill at $0.50 input and $2.50 output, which the discount would take to $0.30 and $1.50. Second, the discount is described as applying to tokens. Commercially, Claude Enterprise costs $20 per seat per month, billed annually with a 20-seat minimum, plus usage on top. Neither Anthropic's post nor the GSA listing says what happens to the seat fee or the seat floor under the OneGov terms, so a small agency should get that answer in writing before forecasting.

A worked example shows how the meter behaves. An analyst who sends 1 million input tokens and receives 200,000 output tokens a month on Sonnet 5.5 would pay $2.40 at the discounted rate, against $4.00 at list. A developer running Claude Code agents all day can burn through many times that, which is why the first metered month is the one to watch. For context on how the cheapest model behaves at scale, see our Claude Haiku 5.5 pricing analysis.

Anthropic vs OpenAI on the same government schedule

Anthropic is not the first vendor through this door. GSA announced its expanded OneGov agreement with OpenAI on September 10, 2026, seven weeks earlier, offering 50% off token-based ChatGPT usage with no platform fee and no minimums. Side by side, from the OneGov listing and each vendor's announcement:

TermAnthropic (Claude)OpenAI (ChatGPT)
Discount40% off tokens50% off tokens
Commitment / minimum$0 / $0$0 / $0
Platform feeNot stated on the listingNone, per GSA
EligibilityFederal, state, local, tribal; Anthropic adds territorialFederal, state, local, tribal
Offer runs toJanuary 31, 2029December 31, 2028
Old $1 offer endedOctober 31, 2026September 30, 2026

Do not stop at 40 versus 50. A discount is a percentage of a list price, and the two vendors' list prices, token counts per task and model mixes all differ. The comparison that matters is cost per completed task on your own workload, at both discounts. Run the same 20 to 50 real tasks through each model, record the tokens, multiply by each vendor's discounted price, and decide on that number.

One more wrinkle sits on OpenAI's side of the ledger: Nextgov reported that OpenAI's blog cites a standard licensing fee of $15 per user per month, without saying whether the 50% discount applies to it. Ask both vendors the seat-fee question in writing.

Per GSA's listing, these are the only two AI model offers on OneGov that extend to state and local entities. Google's Gemini for Government, Microsoft's Copilot offer and AWS's credits are federal-only. A state CIO who wants a discounted frontier model on a federal schedule now has exactly two choices.

The Pentagon wrinkle

The GSA deal does not resolve Anthropic's separate fight with the Defense Department. A US appeals court upheld the Pentagon's "national security supply-chain risk" designation of Anthropic on September 25, 2026, and Anthropic sought a rehearing on October 13, according to Anthropic's own account of the case. A separate designation was vacated on August 27, 2026, when San Francisco federal judge Rita Lin ruled the Pentagon's action under 10 U.S.C. § 3252 was unlawful retaliation against Anthropic's public AI-safety advocacy. The new agreement runs through civilian procurement channels and leaves the military dispute untouched. Any agency planning to standardize on Claude should brief its counsel on the designation first.

Pro tips and common mistakes

The shift from a flat $1 to metered billing is where the real risk sits. Under the pilot, usage had no budget consequence, so nobody tracked it. From November 1, every token has a price, and an agency's appropriation does not flex with a usage spike. There is no cap on the meter unless you set one.

  • Set a not-to-exceed amount before November 1. Issue API keys per program or office so spend can be attributed, route traffic through a gateway that enforces a monthly ceiling, and raise the ceiling once October's usage is priced.
  • Get the rollover terms in writing. Confirm with your contracting officer whether existing Claude access lapses tonight or converts to the consumption terms automatically, and on which vehicle.
  • Ask about the seat fee in writing. The 40% discount is described as applying to tokens. Whether it touches the $20 Enterprise seat or the 20-seat floor is unanswered, and the answer changes a small agency's bill completely.
  • Cloud buyers face a timing choice. Direct purchases qualify now; Bedrock and Vertex AI purchases follow in the coming months. Match the authorization to the workload meanwhile: Claude for Government is FedRAMP High authorized and generally available since September 30, while Claude Enterprise is in process for FedRAMP Moderate through the FedRAMP 20x program.
  • Do not pick a vendor on the discount percentage. Forty percent off a higher list price can cost more than 50% off a lower one. Benchmark cost per completed task, not the headline number.
  • State and local buyers should check cooperative purchasing rules. Eligibility on the GSA listing does not automatically mean your state lets you buy through the federal schedule.

FAQ

When does Anthropic's $1 Claude deal actually end?

GSA's OneGov listing marks the $1 offer as expiring October 31, 2026. It was originally due to end September 30 and was extended by a month. Federal agencies move to metered, consumption-based billing on November 1.

Who qualifies for the 40% discount?

US federal agencies across the executive, legislative and judicial branches, plus state, local and tribal governments. Anthropic's announcement adds territorial governments. Buying runs through GSA's Multiple Award Schedule, directly, or through resellers.

Does the discount apply to Claude bought through AWS or Google Cloud?

Not yet. Direct purchases from Anthropic qualify now; the same 40% discount reaches cloud-provider purchases, including Bedrock and Vertex AI, in the coming months. Anthropic has not given a firm date for the cloud rollout.

How does Anthropic's 40% compare with OpenAI's 50%?

OpenAI's OneGov offer is 10 points deeper and started seven weeks earlier, running to December 31, 2028 against Anthropic's January 31, 2029. But the discounts apply to different list prices, so compare cost per completed task on your own workload rather than the percentages.

Is there a minimum spend or a spending cap?

There is no minimum spend and no commitment. But there is also no automatic cap: under metered billing, a pilot that goes viral inside an agency can run up a bill with nothing stopping it, so set a not-to-exceed amount before November 1.

The $1 era was a land grab: GSA and the vendors priced frontier AI at nearly nothing to get it onto government desks. With OpenAI's deal in September and Anthropic's this week, both now charge for what gets used, at a discount, for roughly the next two years. The pilots bought adoption; the metered deals will show whether that adoption survives contact with a budget line. If your agency is on the $1 deal, get the rollover terms and a spend cap in writing today.

We are tracking every move in the AI price war. See how OpenAI and Anthropic cut AI prices up to 80% in four months for the bigger picture, and our Claude Code explainer for the tool behind Anthropic's government pilots.

This is general information, not procurement or legal advice.

Sources

Exchange rates: £0.747 and C$1.397 per US dollar are October 31, 2026 market rates, verified against live market data at publication.

Related articles

10 min read
19 views
edit_note Written by

Ratul Ahamed

Founder & Editor-in-Chief

Ratul Ahamed is the founder and editor of AI Core Updates, an independent site covering AI tools, models and pricing. He is a web developer based in Bangladesh and also runs All In One Tools, a free collection of browser-based utilities. Contact: [email protected]

Read all stories by Ratul Ahamed arrow_forward
verified_user Human-Reviewed Editorial

forum Comments (0)

Leave a Reply

chat_bubble_outline

No comments yet. Be the first to join the conversation!

Weekly AI Digest

Get our weekly AI tools digest

One email a week analyzing new AI tools, model updates, and practical workflows. Tested and summarized with zero fluff. Unsubscribe any time.

bolt
Hands-on AI Tool Reviews
Pricing math, limits, and alternatives.
verified_user
Strict Anti-Spam Policy
1-click unsubscribe at any time.
mark_email_unread

Subscribe to Digest

Sent weekly • No spam

person_outline
mail_outline

No spam ever. You can unsubscribe with a single click at any time.

cookie

Cookie Settings

We use cookies to enhance your browsing experience, serve personalized content, and analyze our traffic. By clicking "Accept All", you consent to our use of cookies. Learn more