Anthropic has launched Claude Gov, a set of custom artificial intelligence models designed exclusively for U.S. national security customers, including defense and intelligence agencies. The models are already deployed by agencies at the highest level of U.S. national security, with access limited to those operating in classified environments. Claude Gov was developed based on direct government feedback to address operational needs such as strategic planning, operational support, intelligence analysis, and threat assessment. Compared to Anthropic's consumer and enterprise models, Claude Gov features enhanced capabilities for handling classified materials, improved understanding of defense and intelligence documents, enhanced multilingual skills, and improved analysis of complex cybersecurity data. The models have looser guardrails for government use, refusing less often when engaging with classified information, and underwent rigorous safety testing. Anthropic prohibits the use of these models for creating or facilitating illegal weapons, disinformation campaigns, or malicious cyber operations, though certain contractual exceptions are made for selected government agencies. Anthropic's offering directly competes with OpenAI's ChatGPT Gov, introduced in January 2025, and is part of the company's broader involvement in the FedStart program and partnerships with Palantir and Amazon Web Services to deliver AI solutions to U.S. defense customers.