The General Services Administration (GSA) expanded its OneGov agreement with Anthropic on Oct. 27, offering federal, state, local, tribal, and territorial government entities a 40% discount on eligible Claude artificial intelligence (AI) model usage.
The agreement includes models in Federal Risk and Authorization Management Program (FedRAMP)-authorized environments for sensitive unclassified work, according to a GSA press release.
Under the new terms, agencies will pay only for the AI capabilities they use, with no minimum purchase requirements. The consumption-based approach ties costs to actual usage rather than the number of users receiving access.
The expanded agreement follows an earlier OneGov offer that provided Claude access for $1 per user, which is scheduled to expire Oct. 31.
In its own announcement, Anthropic estimated the expanded agreement could save the government $1 billion.
The company also reported growing adoption of its AI tools across government. According to Anthropic, approximately 500 federal, state, and local government organizations currently use Claude.
The company said weekly active users of Claude for Government have increased approximately eightfold since February, while usage across federal civilian agencies has grown about 20-fold over the past year.
The expanded agreement with GSA covers Claude Enterprise, Claude for Government, and the Claude application programming interface (API), giving agencies access to AI capabilities for tasks ranging from administrative work to software development.
The new pricing is available to the executive, legislative, and judicial branches of the federal government, as well as state, local, and tribal governments.
Anthropic said agencies can purchase Claude at the discounted rate through GSA’s Multiple Award Schedule. Anthropic said direct purchasing is available now, and cloud-provider purchasing is expected in the coming months.
The company also announced plans to launch a dedicated program within weeks to help government agencies implement Claude.
On the security front, Anthropic said Claude for Government became generally available Sept. 30 with FedRAMP High authorization, allowing agencies to use the platform for sensitive unclassified workloads.
Agencies can also access Claude through Amazon Web Services’ (AWS) Bedrock and Google Cloud’s Vertex AI, both of which offer FedRAMP High-authorized environments. Anthropic said select Claude models on Bedrock are authorized for Department of Defense Impact Level 4 and 5 workloads in AWS GovCloud.
The expanded agreement is part of GSA’s broader OneGov initiative, which uses governmentwide purchasing power to negotiate standardized pricing for commercial technology products.
GSA said the new terms continue OneGov’s shift toward consumption-based pricing that “scales with actual use.”
The agency has pursued similar pricing arrangements with other AI providers, including OpenAI, as it works to make advanced AI capabilities available across government.
According to GSA, the OneGov strategy has generated more than $1.8 billion in savings for the federal government and taxpayers since its inception, including approximately $1.6 billion through its agreements for AI capabilities, which the federal government has dubbed super intelligence.
GSA said it plans to renew OneGov agreements, where appropriate, as existing offers expire, to maintain access to the technology while securing governmentwide pricing.
The agency did not specify an expiration date for the expanded Anthropic agreement.
Agencies can find ordering details through GSA’s OneGov portal.