ع
Anthropic said it received notification from the Department of Commerce that export controls had been lifted on its two most advanced models allowing restoration of access starting Wednesday. The company made the announcement via social media after the models built on its Claude platform were suspended on June 12. External reporting from CNN confirmed the development noting that the firm stated it would begin restoring access to both Fable 5 and Mythos 5 immediately.
The initial restrictions stemmed from US government concerns that the tools could be exploited by hackers to identify weaknesses in computer systems according to Reuters coverage of the June 13 order. Anthropic had disabled the models for all users worldwide to ensure compliance with the directive targeting foreign nationals. The firm previously indicated that authorities had not provided specific details on the risks involved at the time of the ban.
Fable 5 is positioned for consumer use with strong capabilities in deep reasoning and autonomous performance of complex tasks while Mythos 5 targets enterprise and cybersecurity professionals by detecting and exploiting vulnerabilities in code. Both form part of Anthropic’s competitive push against systems such as OpenAI’s ChatGPT and Google’s Gemini the BBC reported. Forbes added that the partial nature of earlier relief focused primarily on Mythos 5 for approved entities.
Politico reported that the Trump administration’s June 26 action cleared access for a select group of more than 100 US companies and agencies to the Mythos 5 model de-escalating an earlier confrontation over potential cyberattack risks. NPR noted that the government has asserted greater influence over AI by controlling access to such frontier models with OpenAI separately agreeing to let officials screen users of its latest release. The evolution from full suspension to lifted controls on both models reflects ongoing adjustments in policy.
Anthropic had disagreed with the original rationale stating it understood the government identified a potential jailbreak for Fable 5 but maintained this did not warrant broad recall of a system deployed to hundreds of millions. “However we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people” the company said in its earlier response. Al Jazeera and The Guardian detailed how the firm described the initial order as based on a possible misunderstanding.
The sequence of events began with an export control directive on June 12 that prompted the abrupt disablement according to multiple accounts including NPR and Reuters. No immediate public comment emerged from the Department of Commerce despite requests from outlets such as the BBC. Industry coverage indicates the episode forms part of wider US regulatory scrutiny on advanced AI under the current administration.
