Anthropic shut down Fable 5 and Mythos 5 for all customers following a US Commerce Department export control directive issued just days after launch. The government cited a jailbreak that could pose a national security threat, but Anthropic disagreed, warning that such a standard could halt all new model deployments across the industry.
Anthropic abruptly disabled access to its Fable 5 and Mythos 5 models on Friday night, just days after their launch, to comply with a US Commerce Department export control directive restricting their use outside the United States.
The Trump administration was concerned about a reported jailbreak that bypasses classifier-based safeguards on Fable 5 regarding cybersecurity, chemistry, and biology, deeming it a national security threat. Anthropic stated it had only received verbal evidence of a narrow, non-universal jailbreak that found minor software vulnerabilities, and argued that other public models like GPT-5.5 have similar capabilities.
This incident sets a strong precedent for government regulation of AI model deployments. Anthropic warned that applying such a standard—recalling a commercial model based on a narrow jailbreak—could halt all new model deployments across the industry, highlighting the tension between AI safety and innovation.