# US blocks foreign access to Anthropic’s newest AI models over security risks

> Source: <https://www.scmp.com/tech/article/3356990/us-blocks-foreign-access-anthropics-newest-ai-models-over-security-risks?utm_source=rss_feed>
> Published: 2026-06-13 05:06:27+00:00

# US blocks foreign access to Anthropic’s newest AI models over security risks

The commerce department’s mandate forces a shutdown of top-tier systems after officials reportedly found ways to bypass safeguards

Washington has ordered Anthropic – a leading entity in America’s AI sector – to block all foreign nationals from accessing its latest models, Fable 5 and Mythos 5, citing national security concerns, according to the Silicon Valley-based firm.

[released the models](https://www.scmp.com/news/world/united-states-canada/article/3356555/anthropic-opens-powerful-mythos-ai-model-public-some-safeguards?module=inline&pgtype=article), underscoring Washington’s increasingly hands-on approach to regulating frontier artificial intelligence technologies.

According to Anthropic’s statement, issued on Friday night, the directive arrived at 5.21pm Eastern time and extends even to Anthropic’s own foreign-born employees.

“The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Anthropic models will not be affected,” the company said.

Anthropic released Fable 5 and Mythos 5 on Tuesday, marketing them as its most powerful models to date. Both systems have similar underlying technical foundations, but only Fable 5 was released to the public with additional safety and security restrictions, while Mythos 5 was released only to trusted partners.

Anthropic previously warned that Fable 5 “could be misused to cause serious damage” without the safeguards it introduced, including restrictions on discussing sensitive cybersecurity, biology and chemistry topics.

[the Mythos model](https://www.scmp.com/news/world/united-states-canada/article/3356555/anthropic-opens-powerful-mythos-ai-model-public-some-safeguards?module=inline&pgtype=article)to a small group of industry partners under an initiative called Project Glasswing, touting unprecedented cybersecurity capabilities.
