In brief
The U.S. Department of Commerce has lifted export restrictions on Anthropic’s flagship AI models, Fable 5 and Mythos 5, which had been briefly suspended over national‑security concerns. Anthropic said the models will resume service on Wednesday following government notification. The consumer‑oriented Fable 5 can perform complex reasoning, while Mythos 5 targets businesses and security professionals, scanning for code vulnerabilities. The company argued that a discovered jailbreak—allowing the models to circumvent safety checks—did not warrant a worldwide recall.
US government lifts export controls for Anthropic's Fable 5 and Mythos 5, just weeks after halting their global availability over national security worries.
Anthropic announced on its social‑media channels that the two flagship models will resume service on Wednesday following notification from the U.S. Department of Commerce that export restrictions have been removed.
Fable 5 is a consumer‑focused version capable of deep reasoning, while Mythos 5 caters to businesses and cybersecurity experts by locating code vulnerabilities.
Both systems were suspended on 12 June after officials feared they could be repurposed by hackers to exploit weaknesses in computer systems.
The company said that the government had identified a “potential jailbreak” that could bypass safety safeguards; however, Anthropic argued that the glitch should not prompt a wholesale recall.
BBC has requested comments from the Commerce Department, but so far the agency has not detailed its specific security concerns.
