Anthropic has ceased access to its AI models, Claude Fable 5 and Claude Mythos 5, for all users following a directive from the U.S. government. This decision came as a response to an export control letter from Commerce Secretary Howard Lutnick, which stated that these models would be restricted for use by foreign nationals both within and outside the U.S.
Earlier this week, the company had launched these two new models, which are derived from the Claude Mythos Preview. The Mythos Preview was previously tested by a select group of companies and researchers, leading to significant identification and resolution of security vulnerabilities, including hundreds reported by Mozilla.
Despite the initial plans for public accessibility of Fable 5, the directive has prompted Anthropic to disable both models entirely. The company emphasized that access to its other AI offerings remains unchanged. This regulatory action follows concerns about potential national security risks, particularly after reports of another company claiming to bypass security measures in Mythos.