U.S. Eases Restrictions on Anthropic's 'Mythos 5' Only for Domestic Companies

The U.S. government has eased restrictions on Anthropic's top-tier artificial intelligence (AI) model, "Mythos 5," by allowing its use for select domestic companies and institutions. This move comes about two weeks after the Trump Administration announced export controls on Anthropic’s AI models.

According to major foreign media outlets such as Reuters and The Washington Post on June 27, the U.S. Department of Commerce sent a letter to Anthropic on June 26 (local time), notifying the company that Mythos 5 can be used by specific partners who have met security requirements. It is reported that around 100 U.S. companies and institutions are expected to be included in the list of approved entities. Anthropic stated that it will once again be able to deploy the model to U.S. agencies operating critical national infrastructure.

As a result, Korean companies such as Samsung Electronics, SK hynix, and SK Telecom, which participated in the global AI security collaboration project "Project Glasswing" and previously had early access to "Mythos," the predecessor of Mythos 5, are still restricted from using the model. The U.S. government’s control over cutting-edge AI services is also being applied to other companies such as OpenAI. On this day, OpenAI announced the release of its next-generation model "GPT-5.6," stating that it will initially be available only to select institutions designated by the U.S. government. Industry observers have noted that these U.S. government control measures could impact the country’s leadership in the AI sector.

Meanwhile, Anthropic and the U.S. government are reportedly discussing ways to restore access to "Fable 5," a model with safety features intended for general users. Previously, on June 12, the Trump Administration had completely blocked foreign nationals from accessing Mythos 5 and Fable 5, citing national security concerns. With the latest measures, foreign employees belonging to U.S. companies that have received usage approval are now able to use the model.

This content was produced with the assistance of AI translation services.