On Friday night, June 12, 2026, Anthropic suddenly disabled access to its models, Fable 5 and Mythos 5, due to an emergency directive from the US Commerce Department. This unexpected move affects users worldwide and stems from government worries about national security risks related to the models’ abilities in sensitive scientific and technical areas.
The situation escalated quickly after the company received a formal order mandating strict export controls, which effectively barred the models from being used outside the United States.
In a public statement released Friday night, the company confirmed that shutting down Fable 5 and Mythos 5 was the only immediate way to comply with the government’s directive. While this decision affects the latest versions, users can still access other existing Anthropic models without any issues.
Anthropic models: Regulatory Pressure and Security Concerns
The heart of the issue revolves around claims that the models might have vulnerabilities that escape built-in safety measures. According to Ars Technica, an official from the administration pointed out that the government is particularly worried about a reported jailbreak that bypasses classifier-based safeguards. These safeguards were intended to prevent the AI from gene
Reportedly, the administration requested a temporary halt on releasing these specific models to give the national security apparatus time to “harden” infrastructure against possible exploitation. However, Anthropic argues that the evidence provided by officials is limited. The company mentioned that the government has only shared “verbal evidence” regarding a specific jailbreak, which allegedly involved prompting Fable 5 to analyze a certain codebase for software vulnerabilities.
Impact on Development and Future Access
This incident serves as a sobering reminder of the regulatory challenges that frontier AI developers face as their models become increasingly powerful. The sudden loss of access to Fable 5 and Mythos 5 will likely throw a wrench into the workflows of enterprise partners and developers who had already started integrating these models.
Until the government offers more clarity or lifts the export restrictions, the availability of these high-performance models hangs in uncertainty. Industry analysts are keenly observing whether this directive indicates a broader, more aggressive approach to the export of advanced language models.
With the national security apparatus currently reviewing the underlying software architecture of these tools, users can expect increased scrutiny for all future releases of high-parameter models.
FAQs
Why did Anthropic disable Fable 5 and Mythos 5?
The company disabled the models to comply with a directive from the US Commerce Department, which imposed export controls due to concerns over potential security vulnerabilities.
Are other Anthropic models affected by the shutdown?
No, the restriction currently applies only to Fable 5 and Mythos 5. Other models within the company’s ecosystem are fully operational.
When will Fable 5 and Mythos 5 be available again?
While officials suggest that security hardening could be complete in the next few weeks, there’s no confirmed timeline for when public access will be restored.
Source: Ars Technica Anthropic models
