The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…

As a result of a US government directive, we are suspending access to Claude Fable 5 for all users. You can continue to use all other Claude models. Here’s what this means for you: Across Claude products, new sessions will run on your selected default model or Opus 4.8, and existing Fable 5 sessions will end with an error. On the Claude Platform, requests to Fable 5 will also return an error. Please update your integrations to other Claude models. We know this is a disruption to your workflows; we appreciate your patience and support.

This was completely predictable. Anthropic has spent months marketing its models through fear. Every major release comes with the same playbook. The model is too dangerous. It changes the entire cybersecurity landscape. It could become a national security threat. Governments need to step in. Access has to be restricted. Only a small group of approved organizations should be allowed to use the full capabilities. Now the government took them seriously. Anyone who tested Mythos through Glasswing, or saw what @elder_plinius was able to get out of these models, knows the capabilities were real. Mythos was very good. It made certain attacks faster, cheaper and easier to operationalize. It could iterate more effectively, find vulnerabilities, generate working proof of concepts and close the loop with much less friction than most other models. But Anthropic kept talking about it like they had built a digital nuke. The underlying attack surface was already there. The techniques were already there. The vulnerabilities were already there. Other frontier models could already do a meaningful part of the same work. Mythos improved the speed, the reliability and the level of autonomy. That is a serious capability jump. It still does not justify months of apocalyptic marketing. And trying to shift any responsibility toward Pliny or toward the people who tested the model would be ridiculous. Pliny did what security researchers are supposed to do. He pushed the system, broke the guardrails when possible and showed what the model could actually do once the polished demo layer was removed. If a few jailbreaks were enough to trigger a regulatory crisis, the deployment strategy was already cooked. Anthropic built the entire political and commercial narrative around the idea that its safety layer was the thin line separating a useful product from a strategic threat. Then independent researchers showed how fragile that line really was. That is on Anthropic. They wanted the market to believe Mythos was powerful enough to justify special treatment. They wanted regulators to believe frontier models needed tighter controls. They wanted investors to believe they were sitting on capabilities so advanced that only Anthropic could be trusted to decide who should get access. Now the government decided Anthropic should not get to make that decision by itself. The ban is heavy handed. It will hurt legitimate researchers, security teams and foreign nationals working inside the US. It will also push more people toward open models, local deployments and providers outside the reach of US regulators. But Anthropic cannot act surprised. Dario Amodei spent months turning fear into a marketing strategy, a fundraising strategy and a way to build institutional power around Anthropic. Eventually, someone was going to take the narrative at face value. The government did.