Anthropic offers Claude for Government to US civilian agencies


The Decoder reports enterprise features, usage-based billing and departmental controls, while the Pentagon’s ban remains in place.
Anthropic is offering Claude for Government to US federal and state civilian agencies, according to an October 1 report by The Decoder. Agencies receive the same features as enterprise customers, with controls that let departments limit spending and decide which models staff can use.
The platform had already been in open beta—a testing phase—since July, the report says.
According to The Decoder, agencies pay for what they use, subject to fixed caps. Administrators can set separate budgets and model permissions for each department. The platform also includes audit logs, which record activity, and requires two-person approval for sensitive actions.
The Decoder says the platform runs in a FedRAMP High environment, which it describes as the strictest US cloud security level. The outlet also says chats stay on agency devices.
The reported civilian access does not mean Claude is available to the Pentagon, which The Decoder says continues to ban it. The Pentagon labeled Anthropic a supply-chain risk in March after the company refused to drop its bans on mass surveillance and autonomous weapons, according to the report.
The Decoder also reports that Trump previously ordered all agencies to stop using Claude. It remains unclear how that government-wide removal order applies to the civilian offering.
GTIG also reported more distinct flaws disclosed and exploited in January–August 2026 than in all of 2025, an unequal-period comparison.
2 min

Pillar Security said selecting a model could expose data and credentials accessible to the application. The identified fix was released on June 18, 2026.
2 min

The findings describe access information left public for years, but do not establish whether anyone misused it.
2 min