AI 비교하기AI 사용하기AI 최신정보AI 커뮤니티
Our VisionTermsPrivacyContact

US Government Suspends Access to Anthropic Fable 5

US Government Suspends Access to Anthropic Fable 5

Simon Willison
Sunday, June 14, 2026
  • •US government bans foreign national access to Anthropic's Fable 5 and Mythos 5 models.
  • •Anthropic disabled both models globally following a directive received on June 12, 2026.
  • •Company officials state the jailbreaking concern involves capabilities already present in other public models.
  • •US government bans foreign national access to Anthropic's Fable 5 and Mythos 5 models.
  • •Anthropic disabled both models globally following a directive received on June 12, 2026.
  • •Company officials state the jailbreaking concern involves capabilities already present in other public models.

On June 13, 2026, the US government issued an export control directive requiring Anthropic to suspend all access to its Fable 5 and Mythos 5 AI models for any foreign national. This mandate forces the company to disable these models for all customers globally, including foreign national Anthropic employees, to ensure compliance with national security authorities. Access to other Anthropic models remains unaffected. The company received the directive at 5:21pm ET on June 12, 2026, without specific details regarding the national security concern.

Anthropic reports that the government's action stems from a perceived method of "jailbreaking" (a prompt-based technique to bypass safety filters) the models. After reviewing a government-shared demonstration, Anthropic found that the capability was limited to identifying minor software vulnerabilities, a function already widely available in other public models such as OpenAI's GPT-5.5. The company validated that these tools are commonly used by security defenders to maintain system safety. Following the order, service access for Fable 5 began terminating around 6:59pm Pacific (9:59pm ET) on June 12, 2026, returning error 404 responses for API requests.

On June 13, 2026, the US government issued an export control directive requiring Anthropic to suspend all access to its Fable 5 and Mythos 5 AI models for any foreign national. This mandate forces the company to disable these models for all customers globally, including foreign national Anthropic employees, to ensure compliance with national security authorities. Access to other Anthropic models remains unaffected. The company received the directive at 5:21pm ET on June 12, 2026, without specific details regarding the national security concern.

Anthropic reports that the government's action stems from a perceived method of "jailbreaking" (a prompt-based technique to bypass safety filters) the models. After reviewing a government-shared demonstration, Anthropic found that the capability was limited to identifying minor software vulnerabilities, a function already widely available in other public models such as OpenAI's GPT-5.5. The company validated that these tools are commonly used by security defenders to maintain system safety. Following the order, service access for Fable 5 began terminating around 6:59pm Pacific (9:59pm ET) on June 12, 2026, returning error 404 responses for API requests.

Read original (English)·Jun 13, 2026
#anthropic#fable 5#mythos 5#export control#jailbreaking#ai policy