The United States government has issued an emergency export control directive ordering Anthropic to immediately suspend access to its two most advanced artificial intelligence models: Claude Fable 5 and Claude Mythos 5.
National Security Concerns Trigger Sudden Halt
The sweeping federal directive bars any foreign national—including Anthropic’s own overseas employees—from accessing the newly deployed models. To guarantee compliance with the strict mandate, the company was forced to disable both systems for its entire global customer base.
Access was revoked just 3 days after the official public launch of the frontier models.
While the government did not publish full technical details, officials reportedly discovered a critical bypass method, or “jailbreak,” affecting the public Fable 5 model. Its sister model, Mythos 5, which features fewer safety guardrails and is highly proficient at identifying cybersecurity exploits, had only been distributed to a select group of trusted partners.
Anthropic Pushes Back Against the Mandate
The AI safety startup has openly disputed the severity of the government’s findings. Anthropic engineers reviewed the exploit demonstration and concluded that the vulnerability is relatively basic and already present in competing systems, such as OpenAI’s GPT-5.5.
“If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers,” Anthropic warned in an official statement.
The narrative took a sharper turn when David Sacks, co-chair of the President’s Advisory Council on Science and Technology, publicly criticized the firm’s leadership. Sacks claimed that Anthropic CEO Dario Amodei refused initial requests to patch the vulnerability or voluntarily take the model offline.
“Anthropic prioritized the continued offering of the consumer model over safety. The administration issued the export control reluctantly, surprised that Anthropic did not want to cooperate with a reasonable safety request,” Sacks stated on X.
A History of Friction with Washington
This regulatory clash follows a series of disagreements between the AI developer and federal agencies. Earlier this year, Anthropic declined to sign a government contract that would have integrated its technology into mass surveillance programs and autonomous lethal weapon systems. Following that refusal, the Department of Defense designated the firm as a “supply chain risk.”
Frequently Asked Questions (FAQ)
- Which models are affected by the Claude Fable 5 export ban? Only the newly released Claude Fable 5 and Claude Mythos 5 are suspended. Legacy models remain fully operational.
- Why did the government intervene? Officials cited national security risks stemming from a jailbreak vulnerability that could allow bad actors to exploit the models for cyber warfare.
- Will access be restored? Anthropic is currently negotiating with federal regulators to patch the vulnerability and lift the export restrictions as soon as possible.
