U.S. Lifts Export Controls on Anthropic’s Claude Fable 5 and Mythos 5

The U.S. Commerce Department has lifted export controls on Anthropic’s Claude Fable 5 and Claude Mythos 5, ending restrictions that forced the company to suspend access to both models earlier in June.

Anthropic said the export controls were lifted on June 30 and that Fable 5 would become available globally starting Wednesday, July 1, across Claude.ai, the Claude Platform, Claude Code, and Claude Cowork.

The decision follows negotiations between Anthropic and the U.S. government over safeguards for advanced AI systems, especially around jailbreak risks and cybersecurity-related use. Reuters reported that the Commerce Department removed the controls after Anthropic implemented new safety measures and agreed to closer government coordination.

What Changed For Claude Fable 5 and Mythos 5

The restrictions began on June 12, when the U.S. government ordered Anthropic to suspend access to Fable 5 and Mythos 5 for foreign nationals. The order applied to foreign nationals inside and outside the United States, including foreign-national Anthropic employees.

In its earlier June 12 statement, Anthropic said it had no reliable way to verify nationality in real time. As a result, the company disabled both models for all customers to make sure it complied with the directive.

That suspension has now been rolled back. However, access is not identical for both models. Fable 5 is returning broadly, while Mythos 5 remains tied to approved organizations and the company’s Glasswing cybersecurity program.

Model	Status after June 30 decision	Availability notes
Claude Fable 5	Export controls lifted	Global access begins July 1 across Anthropic’s main products and platform
Claude Mythos 5	Export controls lifted	Access restored first to approved U.S. organizations, with expansion planned for Glasswing partners
Other Claude models	Not affected by the June 12 suspension	Anthropic said access to other models was not affected by the directive

Why The U.S. Imposed The Restrictions

The June 12 order followed U.S. government concern over a reported method of bypassing Fable 5’s safeguards. Anthropic said the government believed it had become aware of a jailbreak technique involving cybersecurity tasks.

According to the company’s access suspension statement, Anthropic reviewed a demonstration involving a small number of previously known vulnerabilities. The company argued that the vulnerabilities appeared simple and that other publicly available models could identify them without a bypass.

The dispute still caused a major access shutdown because the order took effect immediately. Anthropic said it had to disable both models broadly because it could not separate users by nationality quickly enough.

• The U.S. order applied to Fable 5 and Mythos 5 access by foreign nationals.
• Anthropic suspended both models for all customers to ensure compliance.
• The reported concern involved a Fable 5 jailbreak related to cybersecurity tasks.
• Mythos 5 had already been limited to trusted cybersecurity partners.
• Fable 5 was designed for broader use with stronger safeguards.

Commerce Lifted The Controls After New Commitments

A letter from Commerce Secretary Howard Lutnick said a license is no longer required for the export, reexport, or in-country transfer of the Mythos or Fable models, according to WIRED.

The same report said Anthropic agreed to proactively detect and address model security risks, work with the U.S. government on protocols and standards for Mythos, Fable, and future models, and strengthen safeguards against jailbreaks.

Commerce also reserved the right to revisit the decision if conditions change or if Anthropic fails to meet its commitments. That means the restrictions have been lifted, but the episode may shape future U.S. oversight of frontier AI launches.

Fable 5 Returns Globally First

Fable 5 is the model most users will notice first. Anthropic’s redeployment notice says Fable 5 will be available globally on July 1 through Claude Platform, Claude.ai, Claude Code, and Claude Cowork.

For Pro, Max, Team, and select Enterprise plans, Anthropic says Fable 5 will count for up to 50% of weekly usage limits through July 7. After that, it will be available through usage credits.

The company also said it would re-enable Fable 5 access on AWS, Google Cloud, and Microsoft Foundry as quickly as possible. That means availability through cloud partners may not appear at the exact same time as availability in Anthropic’s own products.

Access route	Fable 5 status	What users should expect
Claude.ai	Returning July 1	Availability for supported plans as rollout resumes
Claude Platform	Returning July 1	Developers can regain access through Anthropic’s platform
Claude Code and Claude Cowork	Returning July 1	Access resumes across Anthropic’s agent and work products
AWS, Google Cloud, and Microsoft Foundry	Being restored	Cloud partner availability may take longer

Mythos 5 Remains More Restricted

Mythos 5 is not returning as a general consumer model. Anthropic said Mythos 5 access had already been restored for a set of U.S. organizations after government approval on June 26.

The company is still working with the government to expand Mythos 5 access to more domestic and international partners in the Glasswing program. That matters because Mythos 5 has fewer safeguards than Fable 5 and is designed for defensive cybersecurity work.

Reuters reported that Mythos 5 is designed to detect cybersecurity vulnerabilities and had previously been available to a broader set of companies through Glasswing before the restriction.

How The Safeguards Changed

Anthropic said the June 12 directive followed a report from Amazon researchers about a way to bypass Fable 5’s safeguards. The company said it has now added a safeguard to block the behavior described in that report.

When Fable 5 blocks certain cybersecurity requests, Anthropic says some requests may be routed to Claude Opus 4.8 instead. That could frustrate users whose benign security work gets blocked, but Anthropic said the tradeoff helps make Fable 5’s broader capabilities available again.

The company also acknowledged that no AI model can be made fully robust against every possible jailbreak. This admission is important because it frames the issue as risk management, not a one-time technical fix.

• Anthropic added safeguards for the behavior described in the Amazon report.
• Some cybersecurity-related Fable 5 requests may be blocked more often.
• Blocked requests may be routed to Opus 4.8 where appropriate.
• The company is working with Amazon, Microsoft, Google, and Glasswing partners on jailbreak severity standards.
• Anthropic is also expanding pre-release testing and government information sharing.

Why The Case Matters For AI Export Controls

The Fable 5 and Mythos 5 dispute shows how quickly export controls can affect commercial AI access. Unlike traditional software exports, AI model access can involve live cloud products, API users, employees, partners, and customers across many countries.

The order also showed how difficult real-time compliance can be. If a company cannot instantly verify a user’s nationality, it may have to disable access more broadly than the government intended.

WIRED reported that the lifting of controls came after Anthropic reached a deal with the Commerce Department and agreed to stronger cooperation on release protocols and safeguards.

Sonnet 5 Launches During The Same Week

The policy reversal came as Anthropic also launched Claude Sonnet 5, a cheaper and more broadly available model aimed at everyday agentic work, coding, tool use, and professional tasks.

In its Sonnet 5 announcement, Anthropic said the model is available across all plans, serves as the default model for Free and Pro users, and is available to Max, Team, Enterprise, Claude Code, and the Claude Platform users.

Sonnet 5 is not a replacement for Mythos 5 in high-end cybersecurity work. Anthropic said Sonnet 5 has much lower ability to perform cybersecurity tasks than its current Opus models and that Opus 4.8 remains the recommended model for cybersecurity work that needs reduced guardrails.

Model	Primary role	Access pattern
Fable 5	Advanced general model with strong safeguards	Returning globally from July 1
Mythos 5	Advanced defensive cybersecurity model	Limited to approved organizations and Glasswing expansion
Sonnet 5	Agentic work, coding, and general professional use	Available across all plans

What Users And Developers Should Know

Users who lost access to Fable 5 should see availability return through Anthropic’s own products first. Developers using cloud partner platforms should watch AWS, Google Cloud, and Microsoft Foundry availability separately.

Organizations that use Mythos 5 should not assume full public access has returned. The model remains tied to approval workflows and Anthropic’s work with the U.S. government and Glasswing partners.

We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5.

We'll begin restoring access tomorrow, and will share an update soon.

We’re grateful to our users for their patience, and to everyone who worked with us on…

— Anthropic (@AnthropicAI) June 30, 2026

Users moving to Claude Sonnet 5 should also review pricing and model behavior. Anthropic said Sonnet 5 launched with introductory platform pricing through August 31, 2026, before moving to standard pricing.

Bottom Line

The U.S. government has lifted export controls on Claude Fable 5 and Mythos 5, ending an 18-day disruption for Anthropic’s most advanced models. Fable 5 is returning globally first, while Mythos 5 remains focused on approved cybersecurity partners.

The resolution does not mean AI export controls are over. Instead, it creates a template in which advanced model access may depend on stronger safeguards, government cooperation, pre-release testing, and reporting of malicious activity.

For Anthropic, the immediate goal is to restore Fable 5 access while proving that its safeguards can satisfy both users and regulators. For the AI industry, the case is a warning that frontier model launches now sit at the intersection of product strategy, national security, and international access rules.

FAQ