A government order suspended foreign-national access to both models. An Anthropic executive in Seoul says access should return within days.
2026-06-22 · by Moe Ameen
Anthropic disabled access to its frontier models Claude Fable 5 and Claude Mythos 5 in mid-June 2026 to comply with a US government export-control directive. According to Anthropic's public statement and reporting from Bloomberg, CNBC, and CNN, the directive ordered the company to suspend access by any foreign national, whether inside or outside the United States, and the company turned the models off for all customers rather than risk non-compliance. Anthropic has said the directive even applied to its own non-citizen employees.
Fable 5 is the generally available, Mythos-class model Anthropic released on June 9, 2026, with safeguards that route certain sensitive queries to Claude Opus 4.8 instead. Mythos 5 is described as the same underlying model with those safeguards lifted, available only to vetted users through Anthropic's Project Glasswing and trusted-access programs. The directive itself gave no detailed public rationale; reporting has tied the concern to a jailbreak technique against Fable 5, and Anthropic has said its own review of a demonstration found the method surfaced only a small number of previously known, minor vulnerabilities.
Speaking at a press conference in Seoul, Anthropic's Managing Director of International, Chris Ciauri, said the company is "very confident that in the coming days, the models will become available again." As of this writing, restoration of general access had not been independently confirmed, so treat the timeline as a stated expectation rather than a done deal.
If you run your content through Kompozy, this kind of access turbulence does not stop your pipeline. Generation runs server-side on Trigger.dev workers against managed Claude models, and the model layer is abstracted away from your workflow. You write a Persona Brief and approve outputs; you do not hold a raw API key to a specific model that can be switched off underneath you. When the available model changes, the engine keeps generating, scheduling, and publishing across all nine platforms without you re-wiring anything.
There is also a direct content play here. AI access policy is exactly the kind of timely, high-intent topic your audience is searching this week. Drop your take into Kompozy as a source, and it can fan one point of view into a blog post, a carousel explainer, short captioned clips, and platform-native posts, then schedule and publish them across your channels. Being early and clear on a story like this is how a single take becomes a week of content.
As of this writing, Anthropic said it was "very confident" access would return in the coming days, but general restoration had not been independently confirmed. Treat the timeline as the company's stated expectation rather than a completed change.
Per Anthropic, Fable 5 is the generally available Mythos-class model with safeguards that route certain sensitive queries to Claude Opus 4.8. Mythos 5 is described as the same underlying model with those safeguards lifted, available only to vetted users through Project Glasswing and trusted-access programs.
Anthropic said it received a US government export-control directive to suspend access by any foreign national, inside or outside the United States. The directive gave no detailed public rationale; reporting has tied the concern to a jailbreak technique against Fable 5, and Anthropic said its review found the method surfaced only a small number of previously known, minor vulnerabilities.
No. Kompozy runs generation server-side on managed Claude models and abstracts the model layer, so your content pipeline keeps generating, scheduling, and publishing even when raw-API access to a specific model changes.