The Commerce Department lifted its 19-day export block on Anthropic’s Fable 5 on June 30, and the model returns to the Claude Platform, Claude.ai, Claude Code, and Claude Cowork worldwide on July 1. Mythos 5, cleared earlier on June 26 for a set of US organizations, stays domestic. The episode is the first time Washington has pulled a frontier commercial model off global availability and then reversed itself inside three weeks, and the choreography of the reversal matters as much as the outcome.
The sequence: Anthropic shipped the shared architecture behind both models on June 9. On June 12, Commerce issued an export directive citing national security, triggered by an Amazon research report that described prompting Fable 5 to identify software vulnerabilities and, in one instance, generate demonstration exploit code. The directive suspended access “by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.” Anthropic pulled the models the same day.
Commerce Secretary Howard Lutnick said the department “worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government.” Anthropic, for its part, says the Amazon finding didn’t isolate Fable 5 and describes it as “a borderline safeguard case involving routine defensive work” tied to its Project Glasswing defensive-cyber program. “The reported technique did not expose any unique Mythos-level cyber capabilities,” the company said. A new safety classifier, developed jointly with the government, now blocks the specific behavior.
Outside observers aren’t buying the original alarm. Francesco Bailo of the University of Sydney’s AI, Trust and Governance Centre called the jailbreak reports “widely inflated” and said Washington “likely realised it had overreacted.” Tanishq Abraham, formerly of Stability AI, called the reversal a “big deal” and asked the question the industry is quietly asking: “Does the US government need to approve every frontier model release?” Abraham credited Anthropic cofounder Tom Brown with warming the company’s relationship with the administration. It’s worth remembering Anthropic sued the Department of Defense back in March over a “supply chain risk” designation, so the warming is recent.
Investors, per CNBC, spent the 19 days arguing the block handed a scheduling gift to Chinese open-source developers shipping cheaper, near-parity models against Opus 4.8, GPT-5.5, and Kimi K2.7. That pressure, more than any technical finding, is the shape of the incentive Commerce was navigating.
Through July 7, Fable 5 is included in up to 50% of weekly usage limits for Pro, Max, Team, and select Enterprise plans on the Claude Platform, AWS, Google Cloud, and Microsoft Foundry; after that, access shifts to usage credits. The Washington Post’s tech brief noted the process around federal AI intervention “is as opaque as ever.” The precedent set here isn’t that a model came back. It’s that one can be pulled at all, on a research post, inside a weekend.
Sources
- https://www.anthropic.com/news/redeploying-fable-5
- https://www.cnbc.com/2026/06/30/anthropic-says-trump-admin-has-lifted-export-controls-on-claude-fable-5-and-mythos-5.html
- https://www.bloomberg.com/news/videos/2026-07-01/us-lifts-export-restrictions-on-anthropic-s-fable-5-video
- https://www.aljazeera.com/economy/2026/7/1/us-lifts-restrictions-on-powerful-ai-models-fable-mythos-anthropic-says
- https://www.washingtonpost.com/wp-intelligence/ai-tech-brief/2026/07/01/ai-tech-brief-fable-is-back/