US Lifts Block on Claude Mythos 5 for Trusted Institutions

JAKARTA — Claude Mythos 5 can now be released again by Anthropic to more than 100 institutions in the United States after the U.S. government lifted a block that had halted the AI model on Friday, June 27, 2026. The move reopens access for large companies and selected government agencies that had been cut off over concerns about misuse.

The decision eases tensions between the Trump administration and one of the most valuable private players in artificial intelligence. Two weeks earlier, the U.S. imposed export controls on Mythos 5. Anthropic then shut down the model along with its “sibling,” Fable 5, after Amazon and several other companies warned both systems could be jailbroken for harmful use.

Why Claude Mythos 5 matters for the AI industry

The case matters because it touches a question that keeps resurfacing across AI: who gets to use the most powerful models, and under what limits. Once the government steps in, the message is clear. Advanced models are no longer treated like ordinary software. There are security risks, national interests, and reputational stakes for the company behind them.

In a letter to Anthropic chief compute officer Tom Brown, U.S. Commerce Secretary Howard Lutnick wrote that he believed “adequate safeguards” were now in place so that “certain trusted partners” could access Claude Mythos 5. Lutnick also cited “significant progress” in daily intensive talks between the government and Anthropic since the block was imposed. The Verge reported that the letter was sent Friday afternoon.

For Anthropic, the green light keeps its institutional distribution channel open. For the U.S. government, the move shows it still wants room for AI development, but with strict guardrails. No full ban. No free-for-all either. Somewhere between those two poles, the negotiation continues.

What happened to Fable 5

The U.S. letter did not mention Fable 5. Yet the weaker version had previously been the most powerful AI model widely available to consumers before restrictions were imposed. People close to the talks said Fable is also headed for a relaunch, though the timing remains unclear. So the certainty applies only to Mythos 5 for now. The rest is still pending.

The episode shows how quickly an AI model’s fate can change. Open today. Restricted tomorrow. Then discussed again at the government table. For tech companies, that kind of pattern means they must focus not only on model performance, but also on safety, access controls, and usage audits. Without those, even the most capable product can be slammed on the brakes.

Anthropic has long been known for paying close attention to AI safety. Still, the Mythos 5 case shows that a company’s internal standards are not always enough when officials believe the wider risks are bigger. The U.S. government appears to want models like this available only to users considered safe. That is why the phrase used is “trusted partners,” not the general public.

What this means for readers and the AI market

For readers in Indonesia, the story matters not because Mythos 5 is arriving locally right away, but because global AI policy often shapes local markets. If the United States tightens or loosens access to frontier models, companies in Asia — including Indonesia — often adjust strategy too. Service pricing, feature availability, and security standards can all shift in response.

This is also a sign that AI is starting to look more like other strategic industries: there are permits, restrictions, and user screening. Companies that want to adopt advanced models need clear data governance, internal access limits, and usage policies. Without that, even the smartest model can become a new source of trouble.

Anthropic has not disclosed a firm timeline for Fable 5’s release. But the talks described by sources close to the negotiations suggest the door has not fully closed. The market will be watching the next move. And in AI, one government letter can redraw the competitive map in a matter of hours.

Quick summary: the U.S. has lifted the block on Claude Mythos 5; Anthropic may now release it to more than 100 trusted institutions; Fable 5 has not yet been cleared and remains under discussion.

Quick FAQ:
• Who gave the approval? The U.S. Department of Commerce.
• Who can access it? Trusted partners, including large companies and U.S. government agencies.
• Is Fable 5 also approved? It was not named in the official letter.

The next step will depend on follow-up talks between Anthropic and U.S. authorities, while the AI industry waits to see whether Fable 5 follows or stays on hold a little longer.