Anthropic Halts Claude Fable 5 After U.S. Security Order

The artificial‑intelligence company Anthropic has paused the release of its newest model, Claude Fable 5, after U.S. authorities issued an order to restrict access by foreign users. The company said the result was an immediate shut‑down of both Fable 5 and its sibling Mythos 5.

Reports suggest officials were alarmed by a possible jailbreak technique that could bypass the model’s safety measures. Anthropic noted that the vulnerabilities discovered were minor and comparable to those found in publicly available models.

This action follows criticism from the Trump administration, which labelled the technology a supply‑chain risk, and a lawsuit filed by Anthropic against the Pentagon. A federal judge ruled that the Pentagon’s directive could not be enforced, meaning U.S. government agencies can still use the model while the case proceeds.

The company had previously hinted that Fable 5 was too powerful for a public rollout. Critics say the claim is marketing hype, but Anthropic remains committed to proving the model is safe for use after the pause.