Bad News for Fable 5: Anthropic Forced to Pull Two Top AI Models - Here's Why

The U.S. government has ordered Anthropic to immediately suspend access to its two most advanced artificial intelligence models, citing national security concerns over a potential method of bypassing the systems' safety guardrails.

The export control directive, received by Anthropic Thursday at 5:21 p.m. Eastern time, bans all foreign nationals from accessing Fable 5 and Mythos 5, whether inside or outside the United States, including foreign national Anthropic employees. The company said it was forced to disable both models for all customers to ensure compliance. Access to all other Anthropic models is unaffected.

The directive did not specify the national security concern in detail. Anthropic said its understanding is that the government believes it has identified a method of bypassing, or "jailbreaking," Fable 5, essentially prompting the model to read a specific codebase and fix any software flaws.

Anthropic pushed back sharply on the rationale. The company said it reviewed the technique flagged by the government and found it to be a narrow, non-universal vulnerability rather than a broad safety failure, and said it validated that the level of capability demonstrated is widely available from other models, including OpenAI's GPT-5.5, used routinely by cybersecurity professionals.

Prior to Fable 5's launch, Anthropic worked with the U.S. government, the UK AI Safety Institute, multiple private third-party organizations, and internal red teams for thousands of hours. Those tests found Fable's safeguards to be substantially more effective than those of any previously deployed model. No testers identified a universal jailbreak capable of broadly bypassing the model's protections.

Anthropic acknowledged that perfect jailbreak resistance is likely not currently achievable for any model provider, and said it had adopted a defense-in-depth strategy for Fable 5 aimed at making jailbreaks either narrow or very costly to produce, combined with monitoring to quickly detect and shut down successful attacks.

The company said it is complying with the directive but disputes the standard being applied. "We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people," Anthropic said in a statement. "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers."

Anthropic said it believes the government should have authority to block unsafe AI deployments, but called for that process to be "transparent, fair, clear, and grounded in technical facts," adding that the current action "does not adhere to those principles." The company said it would share additional technical details within 24 hours and is working to restore access to both models as soon as possible.