Claude for Government bills agencies with a hard spend cap

The most interesting line in this launch is about money. Claude for Government charges no seat fees: agencies buy usage in fixed increments under a hard not-to-exceed cap, so spending can never go past what an agency has obligated. As of September 30, according to the announcement on the Claude blog, the FedRAMP High platform is generally available to federal and state agencies, after a public beta that started in July.

At a glance

• Federal and state agencies can now contract with Anthropic directly on general-availability terms, and the Claude Code CLI and Claude for Microsoft 365 are entering early access in the same environment.
• Department admins hand prepaid usage down to sub-agencies, while SCIM group mappings set rate limits, dollar caps and allowed models per seat tier, and burndown alerts fire before a balance runs low.
• Anthropic does not say how large the usage increments are or what they cost, so the hard cap tells agencies how spending stops but not how far a budget goes.

If you missed the earlier episodes: according to Claude's beta post, the public beta opened on July 7, 2026, covered Claude Code and Claude Cowork in Claude for Government Desktop, and already described the fixed increments and the not-to-exceed cap. Before that, Unite.AI reports, GSA's OneGov deal of August 12, 2025 offered Claude for Enterprise and Claude for Government to all three branches for a nominal $1. Unite.AI says the current listing describes it as $1 per user, expiring October 31, 2026.

Agencies get Claude's commercial feature set inside a FedRAMP High environment

The platform delivers Claude's coding and agentic work capabilities through a FedRAMP High authorized environment. Anthropic says agencies get capabilities comparable to its commercial customers without compromising compliance requirements, and new capabilities generally arrive on the commercial release cadence. Claude works directly with files on the desktop, so agency staff can use skills, plugins and projects for memo creation, RFP reviews, casework and other tasks.

With Claude Code, public sector teams can build and modernize the software systems that underpin public services. Two more pieces are rolling out in early access through the same environment and under the same administrative controls: the Claude Code command-line interface and Claude for Microsoft 365. Agencies that want into either early access are told to contact Anthropic's public sector team.

No seat fees: agencies prepay in fixed increments under a hard not-to-exceed cap

There are no per-seat fees. Agencies pay for usage in fixed increments with a hard not-to-exceed cap, so spending cannot exceed what an agency has obligated. Administrators define user tiers with spend and model limits per group, track usage by user and by model, and get burndown alerts before a balance runs low.

The admin model follows how departments are organized. Department-level administrators allocate prepaid usage to sub-agencies, and each sub-agency manages its own users. Agencies connect their own identity provider for single sign-on, with self-serve setup in the admin portal, and SCIM group mappings set rate limits, dollar caps and allowed models for each seat tier.

Layered configuration sets defaults for sub-agencies, including what Claude can connect to and which features are available. Procurement officers can contract with Anthropic directly and award on general-availability terms, and agencies do not need a separate cloud-provider relationship to get started.

Sensitive operations on Anthropic's side need two-person approval

Administrative actions are recorded in an audit log that organization administrators can review; Claude's July beta post described that log as hash-chained. Sensitive operations on Anthropic's side require two-person approval. Usage exports contain metering data only, so agencies can answer ATO and inspector general requests without moving sensitive material, and conversation history stays local on the agency-managed device.

Security teams and authorizing officials get audit logs and documentation that supports the ATO process, the authority to operate an agency grants before a system goes live. The FedRAMP Secure Configuration Guide sits in Anthropic's trust center; according to Claude's July post, the change notification and pentest summary are available there under NDA. The app deploys through standard agency MDM platforms, and existing customers can import their conversation history into the desktop application.

How does one spending cap hold across a whole department?

Think of envelope budgeting. The department buys prepaid usage in fixed blocks and splits it into envelopes for its sub-agencies, and no envelope can be overdrawn: once the balance reaches the cap, spending cannot run past it. Burndown alerts are the warning that an envelope is getting thin, sent before it is empty.

Seat tiers are driven by the directory. According to Microsoft, SCIM is a standard 2.0 user-management API that identity systems such as Microsoft Entra ID use to create, update and remove users and groups in SaaS apps. Map a directory group to a tier, and its members get that tier's rate limit, dollar cap and model list.

The data split follows the same logic of keeping things in separate places. Per Claude's July post, inference runs inside the FedRAMP High authorized environment while conversation history sits on the agency's own device, and Anthropic remains the contracted and billing party throughout.

What the announcement leaves out is the price. Anthropic does not disclose how large the fixed increments are or what one costs, so the hard cap tells an agency where spending stops but not how many memos or Claude Code sessions a block buys. In our view, the metering-only usage export is the sharpest design choice here: an agency can answer an inspector general with numbers alone, without copying sensitive conversations into yet another file.

The October 31 OneGov deadline

According to FedScoop, Anthropic's OneGov deal was extended by one month, to Oct. 31, after OpenAI's new OneGov deal moved to discounted, consumption-based pricing until Dec. 31, 2028. Unite.AI reports that GSA is "renewing agreements where appropriate" as OneGov offers expire. Nobody has announced whether Anthropic's $1 offer will be renewed or replaced with usage pricing. No date has been given for the Claude Code CLI or Claude for Microsoft 365 to leave early access either.