Anthropic Restores Access to Mythos and Fable AI After Government OK

Anthropic announced late Tuesday that access to its most advanced AI models, Claude Mythos 5 and Claude Fable 5, would be restored globally on Wednesday, with the US government signing off.

Anthropic reiterated in its statement that Fable 5 had industry-leading cybersecurity safeguards. It also calls for consistent standards for model evaluation and proposes one framework as part of its cybersecurity Project Glasswing. 

In June, Anthropic released Fable 5, the first of its Mythos-level models to be made publicly available. It was only available for a week before the government issued an export control order, commanding Anthropic to block access by any foreign national. Anthropic said the only way to comply was to pull the model for everyone.

"Although we have reached a constructive resolution, these events have made clear that the industry needs a consistent way to assess and fix potential 'jailbreaks' of AI models," the company said. "A shared standard... would help AI developers triage new findings as they arise, launch highly capable models with greater safety, and communicate the level of risk consistently to government and industry partners."

As of 9 a.m. ET on Wednesday, CNET couldn't independently confirm the availability of Fable 5. Access to the model will likely be rolling throughout the day. You can use the new Sonnet 5 model, also released on Tuesday.

Anthropic said it is working to restore access on AWS, Microsoft Foundry and Google Cloud "as quickly as possible." Subscribers on Pro, Max, Team and enterprise plans will be able to use Fable 5 on tasks that use up to 50% of their weekly usage limits (tokens) through July 7, after which Fable will cost extra to use with usage credits. Mythos 5 will only be available to select US companies as part of Project Glasswing.

Commerce Secretary Howard Lutnick, whose office oversees the regulatory Center for AI Standards and Innovation, said on X: "Over the past two weeks, we have worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US government and strengthen America's leadership in AI."

The Trump administration and Anthropic have had a rocky relationship this year. The government had Anthropic declared a supply chain risk this spring after the AI company balked at a Pentagon request that would've allowed Claude AI to be used for all legal military purposes. 

Until the past few months, the administration had taken a hands-off approach to regulating AI. Now, particularly with interventions to block frontier models from public release, it has dramatically changed its stance. An executive order signed on June 2 requires the Defense Department to work with AI companies to have the government review all new frontier models and approve them before they are widely released. Fable and Mythos were blocked under a different legal authority, but showed a similar desire for government approval of public availability for the most capable AI models.