US Government Orders Anthropic to Suspend Access to Fable 5 and Mythos 5 AI Models Anthropic’s most powerful AI models, Fable 5 and Mythos 5, went from record‑breaking launches to a worldwide shutdown in just three days, exposing a deep rift over who controls frontier AI and how far governments will go to contain perceived security risks.

On June 9, Anthropic released Fable 5, a guarded commercial sibling of its ultra‑capable Mythos 5, and it quickly “dominated every benchmark” against OpenAI’s GPT 5.5. The company had already warned that Mythos could pose “serious threats,” keeping it behind a restricted program for vetted cyber‑defenders.

By June 12, that caution collided with Washington’s fears. After Amazon security researchers showed the White House that Fable could, via a prompt like “fix this code,” surface software flaws useful in cyberattacks, Commerce Secretary Howard Lutnick sent Anthropic an export‑control directive, barring access by “any foreign national” and triggering a full shutdown of both models. Anthropic’s statement said the government provided no detailed evidence and that the demonstrated bypass only revealed “previously known, minor vulnerabilities” that “other publicly-available models are able to discover… as well.”

Over the next 24 hours, tense calls between Anthropic CEO Dario Amodei and senior Trump officials failed to produce a voluntary pause, leading to the formal order and a Friday‑night takedown that “successfully forced a tech company to pull its models offline with a swift and unilateral action.” Anthropic complied but called the move “disproportionate,” warning that applying this standard “would essentially halt all new model deployments for all frontier model providers.”

The dispute quickly widened. Cybersecurity leaders, in an open letter signed by nearly 100 experts, argued the ban “has taken the best models away from defenders… without any real risk to justify it,” calling it “not safety, it is sabotage.” Another letter organizer warned the administration had set a precedent that “American models can’t do defensive security research.”

Abroad, allies and competitors interpreted the shock differently. Canadian prime minister Mark Carney likened the shutdown’s systemic risk to the 2008 financial crisis and urged countries to “build out and diversify” AI infrastructure rather than rely on a few U.S. providers. India’s tech leaders seized on the episode to push for a multibillion‑dollar “sovereign AI” fund, warning that depending on American clouds means “your AI infrastructure runs on someone else’s politics.” In Europe, the clampdown “handed one of Europe’s leading AI startups exactly the opening it has been preparing for” as Mistral promotes sovereign, open‑weight models that governments can’t be cut off from overnight.

U.S. officials, for their part, portray the controls as a necessary, if blunt, response to fast‑moving threats. Axios reported that the White House saw Fable and Mythos as national‑security assets and that the models “need to remain locked down until the U.S. government’s national security apparatus is hardened.” Some officials have privately branded Anthropic a “bad actor” whose handling of vulnerabilities amounted to “recklessness” and “screwed us.”

Anthropic insists it supports giving government the power to block unsafe deployments—“as part of a statutory process that is transparent, fair, clear and grounded in technical facts”—but says this action “does not adhere to those principles.” The company has flown senior technical staff to Washington for crisis talks with Commerce in hopes of restoring access and limiting the precedent for future AI releases.

For now, the world’s top‑scoring public AI model is dark, GPT 5.5 is “the strongest model available” largely by default, and governments on both sides of the Atlantic are racing to design new schemes—such as a proposed US‑EU “trusted partner” program—to control who can touch systems like Fable and Mythos at all.

Continue reading https://foxvector.com/stories/019ed2ba-5f5e-0f18-7080-3a0d59ae08ec