Anthropic has halted access to its Fable 5 and Mythos 5 models due to U.S. export controls, citing national security concerns.
Anthropic has officially suspended access to its Fable 5 and Mythos 5 AI models, which were introduced on June 9, following a government directive concerning export controls.
The U.S. government has raised national security concerns, leading to a ban on access to these models for any foreign nationals, including those employed by Anthropic.
The net effect of…— Anthropic (@AnthropicAI) June 13, 2026
The directive, received by Anthropic on June 12 at 5:21 PM Eastern Time, prohibits foreign nationals from accessing the models, resulting in a complete service shutdown.
Anthropic linked the government's action to a technique that allegedly allows bypassing Fable 5's safeguards. They indicated that, after reviewing a demonstration, only a few minor vulnerabilities were discovered, which they believe are easily identifiable by other publicly available models.
Prior to launching Fable 5, extensive testing was conducted by U.S. government agencies, the UK’s AISI, and various third-party organizations, amounting to thousands of hours to validate the model's protective measures. Anthropic asserted that a universal exploit for the model has not been found.
While Fable 5 includes a 30-day data retention policy for clients to examine any potential bypasses, the company noted it had only received verbal indications from the government about a specific, minor vulnerability. This involved directing the model to analyze a particular codebase and correct software issues.
Anthropic reviewed the report believed to be behind the export directive and concluded that similar capabilities exist in other models, including OpenAI's GPT-5.5. The company expressed disagreement with the decision to withdraw its commercial product over a narrow potential bypass, arguing that such a standard could hinder the introduction of new models across the industry.
Despite the suspension of Fable 5 and Mythos 5, access to Anthropic's other products remains unaffected. The company has apologized to its clients and is seeking to clarify the situation.
Interestingly, in June, Anthropic's CEO Dario Amodei called for stricter regulations on AI, advocating for mandatory safety checks for high-capacity AI models rather than just disclosure requirements.
