Anthropic said Tuesday it will provide its Claude for Government and Claude for Enterprise models to every U.S. federal agency, including Congress and the judiciary, for a nominal fee of $1 per agency for the next year. The arrangement was negotiated through the U.S. General Services Administration, giving civil servants access to the startup’s large-language-model tools for sensitive but unclassified work such as document analysis and regulatory review. “America’s AI leadership requires that our government institutions have access to the most capable, secure AI tools available,” Chief Executive Officer Dario Amodei said in a statement. The offer intensifies a growing price war among leading artificial-intelligence vendors for Washington’s business. OpenAI announced a similar $1 deal for its ChatGPT Enterprise product earlier this month, and a person familiar with the matter told the Financial Times that Google is discussing comparable terms for its Gemini models. The government push follows the Pentagon’s July awards of contracts worth up to $200 million to Anthropic, Google, OpenAI and Elon Musk’s xAI, underscoring how access to federal agencies has become a strategic battleground for AI suppliers.