US Halts Foreign Access to Anthropic's Fable 5 AI Model, Citing Security
Iris Coleman Jun 13, 2026 01:22
The US government has suspended access to Anthropic's Fable 5 and Mythos 5 AI models for foreign nationals, raising questions about AI export controls.
The U.S. government has issued an export control directive suspending all foreign national access to Anthropic's advanced AI models, Fable 5 and Mythos 5, as of June 11, 2026. The order applies globally, including to foreign nationals within the U.S., and reportedly stems from concerns over potential "jailbreaking" vulnerabilities, according to Anthropic's official statement.
This move forces Anthropic to disable access to Fable 5 and Mythos 5 for all users, regardless of nationality, to ensure compliance. The company confirmed that its other AI models remain unaffected, but the abrupt suspension is a significant disruption, particularly as the Fable 5 model launched only days earlier on June 9.
Why the U.S. Government Is Concerned
The directive is rooted in national security concerns tied to "jailbreaking"—methods of bypassing safeguards on AI models to elicit unintended outputs. Anthropic acknowledged reviewing a reported narrow jailbreak, which it characterized as minor and similar to capabilities found in other publicly available models like OpenAI's GPT-5.5. Despite this, the government appears to view the risk as severe enough to justify halting access.
Anthropic has defended its safeguards, emphasizing that no universal jailbreaks have been found for Fable 5. The company implemented a "defense-in-depth" strategy that combines robust safeguards with monitoring and a 30-day data retention policy to mitigate risks. Anthropic argues such vulnerabilities are an industry-wide challenge and that the suspension sets a precedent that could stall new AI model deployments across the sector.
What Are Fable 5 and Mythos 5?
Fable 5, billed as Anthropic's most capable public model, features a 1 million-token context window and supports up to 128,000 output tokens. The model is priced at $10 per million input tokens and $50 per million output tokens—double the cost of its predecessor, Claude Opus 4.8. Mythos 5, while less publicized, is believed to be a higher-tier system with capabilities beyond Fable 5. Both models are positioned for advanced use cases such as software engineering, scientific research, and enterprise applications.
The timing of the government's directive is particularly striking, coming just four days after Fable 5's launch. The model had been subjected to extensive pre-launch testing by government agencies, including the U.K.'s AISI, external auditors, and internal teams, with thousands of hours spent scrutinizing its safeguards.
Implications for AI Export Controls
This suspension highlights the growing tension between AI innovation and national security. As AI models become more sophisticated, their potential dual-use nature—helpful for legitimate purposes but vulnerable to misuse—has drawn increasing regulatory scrutiny. Anthropic’s statement underscores the need for a transparent and fair statutory process for such decisions, a standard the company argues was not met in this case.
For users of Fable 5 and Mythos 5, the immediate impact is clear: loss of access to advanced AI capabilities that were quickly gaining traction in high-end professional markets. For the industry, the directive could signal a future where export controls play an increasingly central role in the deployment and accessibility of cutting-edge AI systems.
Anthropic has stated it is working to restore access and will provide additional updates within the next 24 hours. However, the government’s actions may prompt other AI providers to reevaluate their compliance strategies and export policies, especially for models deemed "frontier" systems.
Image source: Shutterstock.jpg)