Anthropic Suspends Access To AI Model Claude Fable 5 After US Government Orders

Anthropic has reportedly disabled public access to its most powerful AI models, Claude Fable 5 and Claude Mythos 5, just days after their release, triggering widespread debate over AI safety, government oversight, and national security concerns. The sudden suspension came after a directive from US authorities instructing the company to revoke access for foreign nationals, including those working within Anthropic itself, over alleged security risks tied to potential misuse of advanced AI capabilities.

According to statements attributed to Anthropic, the company ultimately chose to suspend access for all users globally in order to comply with the order. The firm described the situation as a “misunderstanding” and said it is working to restore access as soon as possible. However, the government’s export control directive reportedly stemmed from concerns that the models could be vulnerable to a so-called “jailbreak,” a method in which AI safety restrictions are bypassed to elicit restricted or harmful outputs.

Anthropic said the government provided only verbal evidence of a “potential narrow, non-universal jailbreak,” allegedly involving prompts that asked the model to analyze code and identify software vulnerabilities. The company disputed the severity of the concern, stating that similar capabilities exist in other widely deployed AI systems and are already used in cybersecurity defense work. It also indicated that it would release further details after reviewing the underlying report.

Also Read: Asus Zenbook Duo 2026 Features Dual 14-Inch 144Hz OLED Displays With Intel Core Ultra 7 Processor

The company expressed concern over the implications of the directive, warning that if such standards were applied broadly, it could significantly slow or halt the deployment of frontier AI systems across the industry. Anthropic also emphasized that its safety filters are already highly restrictive, noting that some users have previously criticized them for being overly cautious. The situation highlights ongoing tension between innovation in AI development and regulatory efforts aimed at controlling potential misuse.

Claude Mythos 5 had previously been positioned as an extremely advanced model, with Anthropic initially limiting its public release due to fears that its capabilities could be exploited for offensive cyber operations or even biological research misuse. The company had instead rolled out a controlled access program for select organizations involved in defensive cybersecurity research. A modified version of the model was later introduced with additional safeguards, designed to restrict high-risk outputs and reduce potential abuse.

Industry benchmarking reportedly placed Fable 5 among the most capable publicly available AI systems prior to its suspension. The abrupt rollback has raised broader questions about how governments and AI companies will balance technological advancement with security risks, particularly as models become increasingly powerful and dual-use in nature. Both regulatory pressure and technical safety concerns are expected to remain central issues as the development of frontier AI continues.

Also Read: Home Coffee Machines: 5 Key Factors To Consider Before Buying, Plus Top 8 Picks