Anthropic restores global access to Claude Fable 5 and Mythos 5 after US Government clearance

Following an intense two-week regulatory freeze driven by national security anxieties, Anthropic has officially announced the global redeployment of its next-generation artificial intelligence models, Claude Fable 5 and Claude Mythos 5.

The global rollout was greenlit immediately after the U.S. Department of Commerce officially lifted an emergency export control directive that had abruptly halted distribution of the highly capable architectures just days after their initial June debut. In a public statement, Anthropic leadership confirmed that exhaustive technical reviews with federal monitors concluded with a structural agreement, allowing the lab to resume its staggered deployment strategy across both public and enterprise sectors.

Restoring the two-tier frontier framework

The lifting of the federal restriction reactivates Anthropic’s dual-product framework, which serves entirely different operational risk profiles using the same underlying foundational model architecture:

• Claude Fable 5: Engineered for the general public, enterprise knowledge workers, and mainstream developers. Fable 5 includes an aggressive, multi-layered defensive safety stack designed to automatically intercept and filter complex requests linked to high-risk domains like biological synthesis and offensive cyber operations.

• Claude Mythos 5: Stripped of the baseline inference-time safety classifiers to preserve maximum analytical flexibility. Mythos 5 is built exclusively for advanced cyber defense and medical research, allowing trusted operators to deploy the system's raw, unrestricted reasoning capabilities to map zero-day system vulnerabilities and isolate infrastructural security gaps.

The technical path to regulatory clearance

The breakthrough in Washington follows intensive coordination between Anthropic engineers, Commerce Secretary Howard Lutnick, and technical advisors from the U.S. AI Safety Institute. To satisfy federal concerns regarding potential "jailbreak" vulnerabilities - techniques malicious actors use to bypass algorithmic guardrails - Anthropic implemented a series of hardened inference-time safeguards across its cloud routing layers.

Importantly, the resolution has laid the groundwork for a broader, standardized safety pact. Anthropic, alongside core cloud infrastructure partners Amazon Web Services, Google Cloud, and Microsoft, has begun co-authoring a shared industry framework to systematically categorize and report model exploit severity. This deeper federal collaboration introduces formalized, pre-release threat testing and mandates a strict 30-day data retention window for safety monitoring on all incoming traffic utilizing the high-compute Mythos 5 tier.

Immediate availability and compute allocations

With economic restrictions lifted, Fable 5 is now active across all primary user touchpoints—including the consumer chat platform, Claude Code, and the collaborative Claude Cowork environment. To mitigate initial server strain during the post-ban surge, Anthropic is offering subscription tiers (Pro, Max, and Team) immediate access to Fable 5 for up to 50% of their standard weekly usage allocations through the first week of July, after which the model will transition to a credit-based consumption model.

Concurrently, access to the high-powered Mythos 5 architecture is being steadily restored for vetted defense consortia and enterprise partners enrolled in the domestic Project Glasswing initiative, with wider deployment across regional AWS and Google Cloud data hubs scheduled to complete over the coming days.