Anthropic's Fable 5 & Mythos AI Banned After Amazon's Warning, Report Claims

Anthropic’s latest and most capable AI models, Fable 5 and Mythos, have hit a major roadblock. The new AI model has been pulled from foreign markets after the Trump administration imposed new restrictions over security concerns reportedly raised by Amazon.

According to a report by The Wall Street Journal, Amazon CEO Andy Jassy alerted senior US officials after company researchers found ways to prompt Fable 5 into revealing information that could potentially aid cyberattacks. To catch you up, Microsoft also recently restricted access to Fable 5 over Anthropic’s data retention policy.

Amazon’s findings reportedly accelerated the ban

People familiar with the matter told The Wall Street Journal that Amazon researchers used a series of prompts to make Fable 5 surface security vulnerabilities that should have remained off-limits.

Jassy reportedly shared those findings with officials including Treasury Secretary Scott Bessent, prompting an urgent response inside the White House. Security experts were then asked to validate Amazon’s claims before the administration moved forward with export restrictions.

The Commerce Department ultimately restricted access to Fable 5 and Anthropic’s unreleased Mythos model for foreign governments, companies, and individuals. Shortly after the restrictions were announced, Anthropic shut down access to both models globally to ensure compliance.

“As a leading cloud provider that serves a large number of private and public sector customers, it’s not uncommon for governments to seek our counsel on potential security risks,” an Amazon spokesperson said. “When they occur, we don’t share the details of these discussions.”

Anthropic pushes back as AI tensions escalate

Anthropic has argued that the vulnerabilities identified by Amazon were relatively basic and could already be discovered using publicly available models. The company also maintains that its safeguards successfully prevented researchers from generating working exploit code.

That being said, the restrictions mark a major escalation in how the US government oversees advanced AI systems. Officials reportedly viewed the export controls as the fastest way to limit potential misuse while Anthropic addresses the concerns.

The administration’s “hope now is that Anthropic remediates the safety issue, the export control is lifted, and Fable goes back into general release,” White House AI adviser David Sacks wrote on X.

For Anthropic, the timing couldn’t be worse. The company is reportedly exploring an initial public offering later this year, while rivals such as OpenAI continue expanding their own cybersecurity-focused AI tools. Not to forget, OpenAI has recently filed for IPO confidentially.