The Commerce Department’s decision to let Anthropic redeploy its Mythos 5 cybersecurity model to a small group of trusted partners ends a two-week standoff. But the restrictions that remain on a related model, Fable 5, signal the government is not done policing advanced AI. Fable 5 stays locked down.
The earlier order bars it from answering certain queries, including those touching cybersecurity and biology. Anthropic’s Claude chatbot now reroutes those questions through other channels.
The Commerce Department’s recent letter, according to multiple reports, makes no mention of easing those limits. That leaves a major piece of Anthropic’s product lineup under direct federal control. The speed of the resolution matters.
Two weeks ago, the government abruptly ordered Anthropic to cut off foreign nationals from Mythos 5 and Fable 5. The concern was straightforward: security guardrails could be circumvented.
Anthropic responded by disabling global access entirely. Then came intensive talks. Commerce Secretary Howard Lutnick wrote to Anthropic’s chief compute officer that the company’s work with the government “have yielded significant progress.” That cleared the way for Mythos 5 to return, but only for a narrow set of users.
Who gets access now? Cyber defenders and infrastructure providers. A small group.
Anthropic said it is working to provision those approved providers and restore their access as quickly as possible. The company also expressed optimism about expanding access to Mythos 5 and eventually making Fable 5 available for general use again.
But that is a future goal, not a current reality. The episode shows how the US government is moving fast to shape the AI industry. Benno Kass, Commerce Department spokesman, put it plainly: “In just two weeks, we have worked diligently to ensure America remains the global leader in AI while safeguarding our security.” The message is that Washington can act, and act quickly, when it sees a risk.
For Anthropic, the immediate effect is operational. It can serve some customers again.
But the uncertainty around Fable 5 hangs over its product roadmap. The company now has to navigate a split reality — one model cleared for limited use, another still muzzled. That is not a normal business environment.
For the broader AI sector, the implications are wider. This is the first major instance of the Commerce Department directly restricting a specific AI model’s deployment. Other companies are watching.
If a leading cybersecurity model can be pulled and then partially restored on government terms, the same could happen to others. The precedent is set.
What comes next depends on the talks. Anthropic said it will work with the government to expand access. Fable 5 remains the test case.
If the company can show that its safety measures are sufficient, the restrictions could lift. If not, the model stays under lock.
The government, for now, holds the key.
