Overview

• Anthropic officially launched its bespoke Claude Gov models on June 5, 2025, marking its first direct AI offering tailored to defense agencies.
• The models are deployed at the highest classified levels and fine-tuned for strategic planning, operational support, and intelligence analysis.
• Claude Gov features improved handling of classified materials and significantly reduced refusal rates when processing sensitive information.
• Although they underwent the same rigorous safety testing as other Claude versions, safety guardrails have been dialed back for government use.
• The announcement heightens competition among AI leaders such as OpenAI, Meta, and Google for multi-billion-dollar national security contracts.