The U.S. Department of Commerce lifted its export controls on Claude Fable 5 and Mythos 5 on June 30, and Anthropic will restore global access to Fable 5 on July 1 across Claude.ai, the Claude Platform, Claude Code and Claude Cowork, with AWS, Google Cloud and Microsoft Foundry re-enabled “as quickly as possible.” The reversal ends a 19-day standoff that started when a national-security directive forced Anthropic to pull both frontier models offline for every user worldwide.

The original June 12 order required Anthropic to cut off both models for any foreign national inside or outside the United States. Because the company couldn’t verify nationality in real time, it suspended access for everyone. The trigger, per Anthropic, was an Amazon report describing a prompt that bypassed a Fable 5 safeguard to identify software vulnerabilities and, in one case, produce exploit code. Anthropic’s own testing found weaker models, including Claude Opus 4.8, GPT-5.5 and Kimi K2.7, could reproduce the same behavior.

Anthropic responded with a new safety classifier that it says blocks the reported technique in more than 99 percent of cases, routing flagged requests to Opus 4.8. The Commerce Department’s Center for AI Standards and Innovation evaluated the updated safeguards. Mythos 5 had already been partially restored on June 26 for roughly 100 U.S. organizations.

Commerce Secretary Howard Lutnick, who authorized the reversal, posted on X that his office “worked closely with Anthropic to analyze and approve Fable 5.” Pro, Max, Team and select enterprise subscribers get up to 50 percent of weekly usage limits through July 7, after which access moves to usage credits.

The lasting artifact isn’t the classifier. It’s the deal underneath it: Anthropic has committed to pre-release government testing of future frontier models and is drafting a jailbreak-severity framework with Amazon, Microsoft and Google that scores incidents on capability gain, breadth, ease of weaponization and discoverability. A private lab, three hyperscalers and Commerce have quietly built the review pipeline that federal AI policy spent two years failing to legislate.

Sources