Anthropic makes Claude for Government generally available to US federal and state agencies

Anthropic announced on 30 September: “Today, Claude for Government is generally available for federal and state agencies.” The service had been in public beta since July. It delivers Claude's coding and agent tools through a FedRAMP High authorized environment. FedRAMP High is the US government's strictest cloud security baseline for sensitive unclassified data. Agencies do not need their own cloud-provider contract to start.

There are no seat fees. Agencies prepay usage in fixed increments, a hard not-to-exceed cap keeps spending within what an agency has committed, and administrators get alerts before a balance runs low.

On controls, department administrators can pass prepaid usage down to sub-agencies. Agencies can connect their own identity provider for single sign-on and set rate limits, dollar caps and allowed models by group. Audit logs record administrative actions, and sensitive operations on Anthropic's side need two-person approval. Conversation history stays on the agency-managed device, Anthropic says.

Two more products are entering early access in the same environment: the Claude Code command-line tool and Claude for Microsoft 365. Anthropic's post reads: “Claude Code CLI and Claude for Microsoft 365 also now available in early access.”

The launch is a commercial push into US public-sector AI, a market where rival labs sell their own government offerings; Anthropic's post names none of them and gives no customer numbers.

• Claimed Claude for Government is generally available for federal and state agencies, after a public beta that began in July. Anthropic
• Claimed It runs in a FedRAMP High authorized environment and is billed by usage, with a hard not-to-exceed cap and no seat fees. Anthropic
• Claimed Claude Code CLI and Claude for Microsoft 365 are in early access in the same environment. Anthropic