Claude for Government Prices Usage With a Hard Cap
Anthropic moved Claude for Government to general availability on September 30, 2026, and the commercial model is the story: no seat fees. Agencies pay for usage in fixed increments with a hard not-to-exceed spending cap. The service runs in a FedRAMP High authorized environment and exits a public beta that began in July. Per Anthropic’s announcement, Claude Code CLI and Claude for Microsoft 365 are rolling out in early access inside the same environment.
The controls are what regulated buyers actually ask about. Administrators set configuration defaults, split spending across departments and define user tiers with spend and model limits. Department-level administrators manage prepaid allocations for sub-agencies. Administrative actions go to audit logs, sensitive operations on Anthropic’s side require two-person approval, and conversation history stays on agency-managed devices. Usage exports carry metering data only, built to support ATO and inspector general requests. Agencies connect their own identity providers with SCIM group mappings.
The cap is the idea worth borrowing. Anthropic’s commercial admin update in July already pointed the same way, adding spend alerts at 75% and 90% of an organization limit, usage and cost broken out by SCIM group and by user, and a Claude Code Value tab with visible, adjustable formulas. Government takes the same instrument and makes the ceiling hard.
For anyone selling AI work into a regulated organization, this is a template. Budget is a ceiling, not a forecast. Spend is allocated to the department that owns the outcome. Auditors get metering, not transcripts. A proposal that arrives with those three answers already written clears procurement faster than one that arrives with a demo.
Before the next engagement, write down who holds the cap, who can raise it, and what the audit log will show.