Why did Anthropic suspend the Mythos and Fable 5 AI models?

Anthropic suspended global access to Mythos and Fable 5 after the US government imposed export controls over national security concerns. Officials reportedly feared that a China-linked group may have gained access to Mythos, potentially allowing the model's capabilities to be replicated.

What is the Mythos AI model?

Mythos is Anthropic’s advanced cybersecurity AI model designed to autonomously identify software vulnerabilities, reproduce exploits and detect previously unknown security flaws. Anthropic considered the model too powerful for general public release and restricted access to approved organisations.

What is Fable 5 and how does it differ from Mythos?

Fable 5 is a consumer-facing version of Mythos that includes safeguards intended to limit advanced cyber capabilities. While based on similar technology, it was designed with restrictions to reduce the risk of misuse.

What does "distillation" mean in the AI industry?

Distillation is a process where developers use a powerful AI model to train another model that reproduces similar capabilities. US officials reportedly feared that access to Mythos could allow foreign actors to reverse-engineer its cybersecurity expertise.

Why are US officials concerned about alleged China-linked access?

Officials worry that advanced cybersecurity AI could be used to strengthen offensive cyber capabilities or accelerate domestic AI development. According to reports, concerns focused on the possibility that Mythos could be copied and redeployed outside US control.

How powerful is Anthropic’s Mythos model?

According to Anthropic, Mythos could autonomously discover high-severity software vulnerabilities, reproduce complex exploits and identify security flaws across major operating systems. The company said it achieved 83.1% accuracy on the CyberGym vulnerability reproduction benchmark.

Image for illustrative purposes only. Not a real photo.

Anthropic halts Mythos after suspected breach

Written by Isaac FrancisPublished Jun 15, 2026, 12:58 AM

Anthropic has suspended worldwide access to its Mythos and Fable 5 artificial intelligence models after the Trump administration imposed export controls amid concerns that a China-linked group may have gained access to the company’s most advanced cybersecurity technology.

The decision followed a June 13 directive from the US Commerce Department, with Anthropic opting to remove both models globally rather than attempt real-time nationality verification for users accessing the systems.

According to a report by Semafor, US officials feared that if Chinese actors obtained Mythos, they could reverse-engineer the model through a process known as distillation and replicate its advanced cyber capabilities.

Anthropic launched Mythos in April 2026 as a highly restricted cybersecurity model capable of autonomously identifying software vulnerabilities, reproducing exploits and discovering previously unknown security flaws across major operating systems and applications.

A consumer-facing version known as Fable 5 was released earlier this month, but White House AI adviser David Sacks said Anthropic had been warned the model had been jailbroken and added that the administration “values Anthropic’s technical capabilities” while describing the issue as “serious but easily resolved.”

The suspension has implications for cryptocurrency firms because Mythos was used to identify vulnerabilities in blockchain networks, smart contracts, decentralised finance platforms and custody systems, with companies including Coinbase reportedly exploring access through Anthropic’s Project Glasswing security initiative.

Anthropic has not provided a timeline for restoring access to either model, leaving cybersecurity researchers, crypto platforms and enterprise users waiting as tensions over artificial intelligence competition between the United States and China continue to escalate.