The restriction affects all users of these models, while access to other Anthropic models remains unchanged.
Restrictions on the most powerful AI models
Anthropic reported that it received instructions to suspend the use of Fable 5 and Mythos 5 for foreign nationals both within the U.S. and abroad. The company stated that it was forced to disable the models for all customers to comply with the requirements.
Previously, Fable 5 was introduced as the first publicly available version of the Mythos-class models—Anthropic’s new generation of systems with advanced capabilities in programming, handling large volumes of data, and tackling complex intellectual tasks.
The company positioned the Mythos 5 model as a more specialized version for a limited circle of vetted partners, including organizations working in cybersecurity and critical infrastructure.
The Security Controversy
According to Anthropic, the U.S. authorities’ decision stems from concerns regarding the potential circumvention of the model’s security mechanisms—a so-called “jailbreak.”
The company stated that the demonstration it reviewed pertained to a limited scenario and did not showcase unique capabilities absent in other public AI systems.
Anthropic disagreed with the authorities’ approach but indicated that it would comply with the directive. The company believes that the regulation of advanced AI models should be based on a transparent technical risk assessment.
What This Means for the AI Market
The situation surrounding Fable 5 and Mythos 5 shows that competition in the field of artificial intelligence is gradually moving beyond technological rivalry between companies. Access to the most powerful models is also becoming a matter of government regulation, export controls, and international technology policy.
For businesses, this means an increased risk of dependence on specific AI tool providers. Companies using advanced models for software development, process automation, and data analysis must consider not only the technical specifications of these solutions but also regulatory restrictions.
A key question for the market is whether such measures will extend to other advanced AI models and whether this will lead to new rules governing access to the most powerful AI systems.
