U.S. Government Orders Anthropic to Shut Down Fable 5 and Mythos 5 AI Models Over Security Concerns

TLDR

• The U.S. government ordered Anthropic to shut down its two most advanced AI models, Fable 5 and Mythos 5, citing national security concerns.
• The shutdown was triggered by a reported jailbreak method, which Anthropic says is narrow and not a universal threat.
• Anthropic complied with the directive but publicly disputed it, calling the evidence insufficient.
• Pre-IPO shares of Anthropic on Hyperliquid dropped 3.7% to around $1,627 following the news.
• Anthropic said it will release technical details about the jailbreak within 24 hours and believes the order stems from a misunderstanding.

---

The U.S. government ordered Anthropic to shut down its two most powerful AI models on Friday. The move rattled pre-IPO markets and put the company on a collision course with Washington.

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.

The net effect of…

— Anthropic (@AnthropicAI) June 13, 2026

The Shutdown Order

Anthropic received the directive at 5:21 p.m. ET on Friday. The order, issued under national security authorities, required the company to suspend all access to Fable 5 and Mythos 5 for any foreign national, inside or outside the United States.

That included foreign national employees at Anthropic itself.

To ensure compliance, Anthropic pulled both models for all users entirely. Its other models, including Opus 4.8, were not affected.

Fable 5 and Mythos 5 had launched just days earlier. They were built on top of Mythos Preview, a general-purpose model Anthropic said had found thousands of vulnerabilities in critical software.

The government told Anthropic it had learned of a method to jailbreak Fable 5. A jailbreak is a technique used to bypass a model’s safety guardrails.

Anthropic reviewed the method and pushed back. The company said the technique was narrow, not a universal jailbreak, and relied on a small number of previously known, minor vulnerabilities.

It also said other publicly available models, including OpenAI’s GPT-5.5, can find the same vulnerabilities without any bypass method at all.

“To date, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws,” Anthropic said.

The company added that defenders use that exact task every day.

Impact on Pre-IPO Shares

Markets responded quickly. The Anthropic perpetual contract on Hyperliquid, which tracks where traders expect the company’s equity to price, fell 3.7% on Saturday to around $1,627.

The contract had been trading above $1,800 after Fable 5 launched. Open interest sits near $8.6 million, small compared to the SpaceX perp but meaningful for a company that has not yet filed for an IPO.

Anthropic said it believes the order is the result of a misunderstanding and is working to restore access as quickly as possible.

The company plans to share full technical details about the alleged jailbreak within 24 hours.

Anthropic warned that applying this standard broadly “would essentially halt all new model deployments for all frontier model providers.”

The company, which built its reputation on safety-first AI, is now publicly disputing a government directive on the grounds that the evidence does not meet its own stated standard.

Whether the order gets reversed, narrowed, or extended will shape Anthropic’s path to a public listing.