limited Mythos – Anthropic says the US government has approved limited redeployment of its Mythos 5 cybersecurity model to specific US critical-infrastructure organizations—roughly two weeks after an export-control order forced the company to disable access for the model and F
Anthropic said it worked through the weekend with the US government—then received a decision that changed what it could safely ship to customers.
On Friday. the company said the government notified it that Mythos 5. described as its strongest cybersecurity model. can be redeployed to a set of US organizations that operate and defend critical infrastructure. Anthropic said it is restoring access quickly and continuing to press for broader availability of Mythos 5 and for Fable 5 to become available for general use again.
“Since June 12, we’ve been working closely with the US government to restore access to Claude Mythos 5 and Fable 5. Today. the government notified us that Mythos 5. our strongest cybersecurity model. can be redeployed to a set of US organizations that operate and defend critical infrastructure. ” Anthropic said in a statement.
The company added: “We’re restoring access for these organizations quickly, and we’re continuing to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again.”
Two weeks earlier. Anthropic said it had disabled access to both Fable 5 and Mythos 5 in response to a US export-control order. That order barred access to the two models by foreign nationals. including foreign national employees inside Anthropic. which is based in San Francisco. Anthropic at the time said the order was tied to government concerns that safeguards in the models could be bypassed. but the company disputed the severity of the issue. It also said the government did not provide specific details “of its national security concern.”.
The same balancing act is playing out across the frontier AI market. On Friday, Anthropic’s rival, OpenAI, said the US government requested that it limit preview access for its latest series of models, called GPT-5.6, to a “small group of trusted partners.”
OpenAI said such government-imposed limits should not become the norm. “We don’t believe this kind of government access process should become the long-term default,” the company said. “It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them.”
The sequence on Anthropic’s side is stark: access was paused after the export-control order, then partially restored for Mythos 5 to US organizations tied to critical infrastructure. Even with that reversal, Fable 5 remains suspended and the scope for Mythos 5 is still constrained.
For readers tracking how powerful AI is governed. the immediate human impact is clear in the work itself—access changes quickly. permissions narrow and widen. and employees and customers feel the shake-up in real time. Friday’s carveout offers a practical opening. but it also underscores how tightly model availability can hinge on government judgments about security and safeguards.
Anthropic Mythos 5 Fable 5 US export-control order Claude Mythos 5 critical infrastructure cybersecurity model OpenAI GPT-5.6 trusted partners frontier AI regulation
So they turned it off then turned it back on? Cool.
This sounds like the government is basically picking who gets what cybersecurity AI. Like if you’re not “critical infrastructure” then tough luck? Also “two weeks after” the export thing… seems kinda fast but whatever.
I don’t get why a model that helps with security needs export controls. Isn’t cybersecurity stuff already supposed to be for everyone? If it was that dangerous they should’ve just blocked it forever. Sounds like they’re arguing over loopholes.
“Foreign nationals employees inside Anthropic” like… so they only allow US citizens to use it? That seems weirdly political for a tool to defend power grids. I’m sure there’s a reason but it still makes me nervous like what if the safeguards can be bypassed by the same people who get access. Also Mythos 5 sounds like a sci-fi thing so I’m like why do we even trust it.