U.S. lifts export controls on Anthropic's Claude Fable 5 and Mythos 5, ending 19-day shutdown

Anthropic's Claude Fable 5 and Mythos 5 are coming back online. The U.S. Department of Commerce lifted its export controls on both models on June 30, 2026, Anthropic announced on X, ending a 19-day shutdown that froze access for enterprise customers across dozens of global cloud platforms.

Anthropic said it will begin restoring access to Fable 5 for general users on the Claude platform, Claude.AI, and Claude Code starting July 2, according to CNBC. Access through Amazon Web Services, Google Cloud, and Microsoft Foundry will follow as quickly as possible. Mythos 5 access is being restored first to a group of approved U.S. organizations, with broader expansion planned through the company's Glasswing cybersecurity program.

How the shutdown unfolded

Anthropic launched both models on June 9, 2026. Claude Fable 5 went to enterprise customers and paid subscribers, while Claude Mythos 5 was restricted to vetted partners through Project Glasswing, a cybersecurity consortium. As MarketScale reported, the two are built on the same underlying model, with Fable 5 including guardrails that route high-risk cybersecurity queries to a fallback model, triggering in fewer than 5% of sessions on average.

Three days after launch, on June 12, the Department of Commerce issued an export control directive requiring Anthropic to suspend access for any foreign national, including Anthropic's own non-U.S. citizen employees, regardless of location or use case. Anthropic determined that filtering users by nationality across its global cloud infrastructure was not technically or legally feasible on short notice, so it enforced a universal shutdown, pulling Fable 5 and Mythos 5 from AWS Bedrock, Google Cloud, Microsoft Foundry, Snowflake, Box, and the direct Claude APIs simultaneously, according to MarketScale.

The government's stated rationale was a reported jailbreak technique that could circumvent Fable 5's safeguards and expose the more powerful capabilities built into Mythos 5. Anthropic contested that framing, telling Forbes that the demonstration was used to identify a small number of previously known, minor vulnerabilities, rather than execute a sophisticated cyber breach, and that comparable techniques already existed for other deployed models not subject to similar controls.

The path to restoration

The full lift came after a staged reopening. On June 26, Commerce Secretary Howard Lutnick sent Anthropic a letter granting Mythos 5 access to roughly 100 companies and federal agencies, determining that appropriate safeguards were in place for those trusted partners, as viewed by CNBC. Fable 5 remained suspended at that point, with negotiations continuing.

On June 30, Lutnick posted on X that the government had worked closely with Anthropic over the prior two weeks to analyze and approve Fable 5 and strengthen America's leadership in AI, according to Forbes. White House Chief of Staff Susie Wiles added her own statement, referencing the administration's executive order on advanced AI innovation and security and framing the resolution as part of a broader priority to get the best tech deployed as quickly and safely as possible, as reported by Forbes.

Lutnick's letter also warned that the administration reserves the right to reimpose restrictions if circumstances change or if Anthropic fails to adhere to its commitments, Forbes noted. Anthropic, for its part, committed to proactively detecting and addressing security risks and informing the government if it detects malicious activity involving the models.

Competitive pressure added urgency

The shutdown did not happen in a vacuum. CNBC reported that the export control period coincided with a rapid rise in Chinese open-source models that are proving nearly as capable as leading U.S. offerings at significantly lower cost. A number of tech executives and investors publicly warned that the restrictions were effectively handing Chinese developers valuable time to close the capability gap, adding competitive pressure on top of the legal and regulatory complexity.

What enterprise buyers need to rethink

For enterprise teams, the episode exposed a risk category that most vendor contracts were not designed to address. Finance, healthcare, SaaS, and critical infrastructure customers found core intelligence services disabled without exception, advance warning, or immediate recourse, as MarketScale detailed. Force majeure clauses written before 2026 did not contemplate a government-mandated, instantaneous AI suspension affecting every platform and integration at once.

The shutdown also established what MarketScale described as a new template for frontier model governance: tiered, authorization-based access, with a trusted-partner tier sitting between full public availability and total suspension. Mythos 5's phased return through the Glasswing program is the first concrete example of that structure operating at scale.

Anthropic told users via X that Fable 5 will count for up to 50% of weekly usage limits through July 7 for Pro, Max, Team, and select enterprise plans, according to CNBC. The company said it will continue working with the government to expand Mythos 5 access to additional domestic and international partners through Glasswing.