Anthropic Enhances AI Partnership with US Government, Offering 40% Discount

Anthropic has significantly expanded its partnership with the US government, introducing a 40% discount on token usage for all Claude models under a revised OneGov agreement with the General Services Administration (GSA). This new pricing structure, announced on October 27, 2026, replaces a flat fee with a consumption-based model, allowing agencies to pay based on actual usage.

The agreement now covers federal, state, local, tribal, and territorial governments, broadening the potential customer base for Anthropic's AI products, including Claude Enterprise and Claude for Government. The procurement process has been simplified, enabling agencies to purchase directly or through resellers without a minimum order requirement, and ensuring access to FedRAMP-compliant environments.

This expansion follows a previous arrangement that offered government users access to Claude for a nominal fee of $1, which supported FedRAMP High workloads. The transition to a consumption-based pricing model indicates a shift from evaluation to practical deployment of AI technologies in government operations.

In comparison, OpenAI has also negotiated a deal with the GSA, offering a 50% discount on token usage through December 2028, which places Anthropic's offer in a competitive context. However, the actual costs for agencies will depend on various factors, including the chosen model and token consumption.

The implications for government AI adoption are significant, as the new pricing model reduces barriers to entry and allows for more flexible budgeting. This could lead to increased adoption of AI technologies across various government sectors, while also fostering a competitive environment among AI vendors.