Anthropic has restored access to Claude Fable 5 and Mythos 5 starting July 1, 2026, after the US government lifted export restrictions imposed just weeks earlier. The export control order, issued June 12, required Anthropic to block access to foreign nationals. Since the company couldn’t verify nationality in real time, it suspended access for all users on both models. With the June 30 reversal, the models are now available globally on Claude Platform, Claude.ai, Claude Code, and Claude Cowork.

The restriction lasted less than three weeks but created significant disruption for users and developers relying on the most advanced Claude models. The government’s June 12 order came after Amazon researchers published a report identifying a method to bypass Fable 5’s safety guardrails by prompting the model to identify software vulnerabilities. Anthropic’s testing later showed that less capable models, including Claude Opus 4.8 and competing offerings from other companies, could identify the same vulnerabilities, undermining the rationale for the export control.

The Timeline and Lifting of Restrictions

On June 12, the Commerce Department flagged Fable 5 for cybersecurity concerns and immediately banned international access. Anthropic responded by suspending the model entirely for all users to avoid liability. The blunt approach—blocking everyone rather than just foreign users—reflected the practical impossibility of nationality verification at internet speed.

By June 30, the Commerce Department concluded the export control was unnecessary and revoked the order. The reversal suggests government officials reviewed Anthropic’s findings about similar capabilities in competing models and determined that restricting only Fable 5 created an uneven playing field without meaningful security gains. On July 1, Anthropic flipped the switch back on.

Access and Pricing Changes

For Pro, Max, Team, and select Enterprise plans, Fable 5 is included for up to 50% of weekly usage limits through July 7. After July 7, the model will be available via usage credits, requiring paid add-ons. This tiering suggests Anthropic wants to manage demand while users readjust to the model’s availability.

The pricing shift also acknowledges that Fable 5 costs more to operate than Claude Opus 4.8. By capping free access, Anthropic can distribute compute capacity while pushing power users toward paid tiers. The July 7 deadline gives existing free users a week to experience Fable 5 before paywalling access.

What This Reveals About AI Regulation

The episode exposed gaps in how government restricts AI capabilities. When one model is restricted but others offer similar functionality, the restriction mostly inconveniences users of that specific vendor. It doesn’t solve the underlying security concern—it just redistributes users to competitors.

The reversal also shows government agencies listening to technical feedback. Anthropic provided data showing that the vulnerability Amazon researchers found was not unique to Fable 5. The government evaluated that information and changed course within weeks. That responsiveness is unusual for federal regulation.

The three-week suspension proved more disruptive to Anthropic’s business than the government intended. But the reversal suggests a pattern: export controls will be imposed quickly and modified based on industry feedback. Companies building frontier AI will need agility to adapt to rapid regulatory changes.