The U.S. Department of Commerce has officially lifted export controls on Anthropic’s Claude Fable 5 and Claude Mythos 5 models, authorizing the company to restore global access starting tomorrow. Commerce Secretary Howard Lutnick and White House Chief of Staff Susie Wiles confirmed the decision, describing it as the outcome of unprecedented coordination between the federal government and private AI developers under a strategy designed to preserve U.S. technological leadership. Originally deployed on June 9, the models share identical foundational architectures but feature distinct security configurations. Fable 5 is engineered for general users with stringent protective layers, while Mythos 5 operates with moderated constraints for vetted cybersecurity firms and critical infrastructure operators. Just three days after launch, the Commerce Department imposed a comprehensive access ban citing national security risks related to potential jailbreak exploits that could reveal software vulnerabilities. Unable to reliably isolate foreign users across its global infrastructure, Anthropic temporarily suspended access worldwide. The company challenged the regulatory rationale, noting that the documented exploits addressed known, low risk issues and that comparable findings are standard across other frontier models. Enforcement gradually shifted following June 26, when authorities permitted Mythos 5 to resume limited access for approved institutions and international staff, leaving Fable 5 under restriction. The full rescission of export controls now enables Anthropic to resume broad deployment across all regions and subscription tiers, though exact implementation timelines and updated verification protocols remain pending. The policy shift coincides with the release of two new systems. Anthropic introduced Claude Sonnet 5, its most advanced agent model to date, featuring autonomous task planning, browser automation, and terminal integration. Priced at two dollars per million input tokens and ten dollars per million output tokens through September before adjusting to three and fifteen dollars respectively, Sonnet 5 exhibits markedly reduced offensive cybersecurity capabilities compared to Opus 4.8 and Mythos 5. The company also launched Claude Science, a specialized research environment that consolidates literature databases, Jupyter notebooks, R environments, and high performance computing clusters. The platform allows researchers to execute literature reviews, generate analysis code, and run simulations through natural language prompts while maintaining auditable, verification backed computational records. The resolution reflects a maturing regulatory framework for advanced AI, demonstrating how national security oversight and commercial innovation can align. Anthropic’s successful navigation of compliance requirements while advancing its product roadmap underscores a growing consensus between Washington and leading AI developers on safe, competitive deployment standards.