Anthropic abruptly disables Fable 5 and Mythos 5 after US export directive
New AI models pulled within days due to national security jailbreak claims.
In a sudden move late Friday night, Anthropic completely disabled access to its newly released Fable 5 and Mythos 5 models. The action came hours after the company received a directive from the US Commerce Department, subjecting the models to export controls that restrict their use outside the United States. Anthropic stated that the only way to ensure immediate compliance was to abruptly shut off access for all customers, while other models remain unaffected. The decision follows reports of a jailbreak that reportedly bypasses classifier-based safeguards designed to block prompts related to cybersecurity, chemistry, and biology.
Anthropic has pushed back against the government’s rationale, arguing that the alleged jailbreak is narrow and non-universal, involving getting Fable 5 to review a specific codebase for minor software vulnerabilities. The company notes that similar capabilities exist in other publicly available models like GPT-5.5. In its announcement, Anthropic warned that if this standard were applied across the industry, it would essentially halt all new model deployments from frontier providers. The company apologized to customers for the disruption, calling it a result of a misunderstanding, and promised more details within 24 hours.
- Fable 5 and Mythos 5 were disabled days after launch following a US Commerce Department export directive
- Government cited a narrow jailbreak allowing code review for vulnerabilities; Anthropic called it minor and non-universal
- Anthropic warns this precedent could halt all frontier model deployments if applied industry-wide
Why It Matters
This sets a potential precedent for government intervention in AI model releases, affecting industry deployment timelines.