Anthropic has suspended access to two of its most advanced AI models — Fable 5 and Mythos 5 — following government security concerns, according to multiple reports. The move comes even as the models sit at the top of global AI performance rankings: according to Jefferies, Fable and Mythos 5 lead the world in model intelligence, with the gap between US and Chinese AI systems now widening to 13%.
The security alarm intensified after The Verge reported that China may have accessed Mythos, a claim that adds urgency to what might otherwise look like a routine regulatory pause. Anthropic has not stood still: according to Digit.in, the company has dispatched senior executives to the US to negotiate the removal of the restrictions and get both models back online.
The suspension is drawing scrutiny beyond Washington. Tech Policy Press notes the episode has exposed the White House's lack of a clear AI safety playbook — leaving businesses and policymakers without a defined framework for how to handle models deemed sensitive.
The fallout is already being felt internationally. According to Livemint, Indian AI startups — most of which build their products on top of foreign models like Anthropic's — are now confronting fresh risks from policy shifts and access restrictions they have no control over.
The episode underscores a structural vulnerability at the heart of the AI economy: when the world's most capable models are controlled by a handful of companies subject to government orders, everyone who built on top of them is exposed overnight.