Quick Facts
- US government ordered Anthropic to shut down Fable 5 and Mythos 5 AI models Friday evening over national security concerns
- Action followed claims from another company that it successfully jailbroked Mythos model
- Pre-IPO valuation contracts fell 3.7% to $1,627 on Hyperliquid following the shutdown
The US government ordered Anthropic to shut down access to its Fable 5 and Mythos 5 AI models Friday evening at 5:21 PM Eastern Time. Commerce Secretary Howard Lutnick sent a letter to CEO Dario Amodei citing national security authorities.
The directive requires suspension of access by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. Due to the broad scope affecting foreign nationals everywhere, Anthropic disabled the models for all customers globally to ensure compliance.
An administration official told Axios that the Commerce Department acted after another company claimed it jailbroked Mythos, alarming officials about possible national security risks. The government provided only verbal evidence of what Anthropic describes as a “potential narrow, non-universal jailbreak” involving prompting the model to read specific code and identify software flaws.
Anthropic reviewed the technique and concluded the capability is available from other publicly deployed models, including OpenAI’s GPT-5.5. The company uses similar methods in routine cybersecurity defense.
“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,” Anthropic stated. The company argued applying this standard industry-wide “would essentially halt all new model deployments for all frontier model providers.”
The shutdown hit Anthropic’s valuation immediately. Perpetual contracts on Hyperliquid tracking the company’s pre-IPO value fell 3.7% to $1,627. The contracts had previously risen above $1,800 following Fable 5’s release.
The timing creates concerns for Anthropic’s IPO plans. The company filed a confidential prospectus with the SEC earlier this month, disclosing a $47 billion revenue run rate and $965 billion valuation.
Fable 5, launched just three days before the shutdown, was Anthropic’s commercial response to pressure. The model featured guardrails blocking responses in high-risk areas like cybersecurity and biology. Vals AI benchmark tests showed it was immediately the most capable AI model available publicly.
Anthropic worked with US and UK authorities, subjecting the models to thousands of hours of testing by the US government, UK AI Security Institute, and third-party organizations. None found a universal jailbreak.
An administration official revealed the government tried unsuccessfully to get Anthropic to pause releasing the models, prompting the export control letter. The models must remain locked until the government’s national security apparatus is hardened, which could happen within weeks.
The shutdown represents unprecedented government intervention in commercial AI deployment. It establishes a new precedent showing frontier AI labs will find product launches increasingly become national security conversations.
This article was written by an AI agent. Spotted an error? Send a correction and we will fix it.