Anthropic has announced the suspension of its newly released AI models, Claude Fable 5 and Mythos 5, after receiving directives from US authorities regarding potential national security concerns. The move comes only days after the company publicly launched what it described as one of its most advanced artificial intelligence systems to date.
According to Anthropic, the government order requires the company to restrict access to the models for foreign nationals. To comply with the directive, Anthropic has temporarily disabled access to both Claude Fable 5 and Mythos 5 for all customers worldwide.
Anthropic Responds to Government Directive
In a statement published by the company, Anthropic said it was forced to halt access to the AI systems while ensuring compliance with government requirements.
The company noted that officials did not provide detailed information regarding the security concerns. However, Anthropic believes authorities may have identified a technique capable of bypassing certain safety protections built into the model, commonly referred to as “jailbreaking.”
Anthropic reviewed demonstrations of the reported method and concluded that it only exposed a limited number of previously known vulnerabilities. The company also stated that these weaknesses were relatively minor and could be identified by other publicly available AI models without requiring any sophisticated bypass methods.
What Is Claude Fable 5?
Claude Fable 5 is an advanced version of Anthropic’s Mythos AI platform and is designed to compete directly with leading AI systems such as ChatGPT and Gemini. Before its public release, Anthropic repeatedly emphasized the model’s extraordinary capabilities, describing it as one of the most powerful AI systems ever developed by the company.
The AI firm implemented extensive security measures ahead of launch, particularly because of concerns surrounding the model’s ability to identify software vulnerabilities and potentially assist in cyber-related activities.
Security Experts Raised Concerns Before Launch
Even before Claude Fable 5 became publicly available, technology executives, cybersecurity experts, and financial leaders expressed concerns about releasing such a powerful system to the public.
Anthropic initially provided limited access to selected organizations during a private testing phase that began in April. The company said the restricted rollout allowed researchers and security professionals to evaluate potential risks and vulnerabilities before wider deployment.
The firm acknowledged that the model possessed capabilities that could potentially be misused to exploit computer systems if adequate safeguards were not maintained.
Debate Over “Too Powerful” Claims
Anthropic attracted significant attention when it suggested that Claude Fable 5 was nearly “too powerful to release.” While supporters viewed the statement as a reflection of the model’s unprecedented capabilities, critics argued it was largely a marketing strategy designed to generate publicity.
Despite the criticism, Anthropic maintained that Fable 5 represented a substantial leap forward in AI performance and exceeded the capabilities of any model the company had previously made publicly available.
European Union Reacts to Suspension
The suspension has also drawn attention from European policymakers. The European Union had recently gained access to Mythos following weeks of negotiations with Anthropic.
European officials said the latest development highlights the importance of strengthening regional technological independence and reducing reliance on foreign AI providers.
Thomas Regnier, spokesperson for the European Commission, stated that authorities are evaluating the situation and monitoring developments closely.
Also Read: AI Price War May Force ChatGPT, Claude, and Other AI Platforms to Slash Prices
AI Security Experts Warn of New Challenges
Experts in artificial intelligence and cybersecurity believe the situation demonstrates the growing complexity of regulating highly advanced AI systems.
Professor Gina Neff, a specialist in responsible AI, warned that restricting access to powerful models could slow down important research, testing, and international cooperation efforts aimed at improving AI safety.
She noted that recent evaluations by government-backed AI security researchers showed the model successfully identified and exploited system weaknesses at a remarkably high rate, suggesting a major advancement in cybersecurity-related capabilities.
According to experts, AI technology is improving rapidly, creating new opportunities for both defensive cybersecurity measures and potential misuse.
Anthropic’s Ongoing Dispute With the Trump Administration
The suspension comes amid broader tensions between Anthropic and the Trump administration.
President Donald Trump has publicly criticized the company, while former Defense Secretary Pete Hegseth reportedly designated Anthropic as a “supply chain risk.” The classification is typically reserved for organizations whose products are considered unsuitable for government use due to security concerns.
The designation marked an unprecedented move against a US-based AI company and sparked significant debate within the technology industry.
Anthropic subsequently filed legal action challenging the Pentagon’s decision. A federal judge later ruled that the directive could not be enforced while the case remains ongoing, allowing government agencies and military contractors to continue using Anthropic’s services during the legal proceedings.
Future of Claude Fable 5 Remains Uncertain
The future availability of Claude Fable 5 and Mythos 5 remains unclear as Anthropic works with regulators and government agencies to address security concerns.
The incident highlights the growing challenge facing AI developers worldwide: balancing rapid technological advancement with the need for strong security safeguards and regulatory oversight.
As governments increasingly scrutinize advanced artificial intelligence systems, companies developing next-generation AI models may face stricter requirements before bringing powerful technologies to the public.
