This is so significant for us here in India.
The US government just told Anthropic to cut off its two most powerful models from foreign nationals, inside or outside America.
Anthropic could not cleanly fence off foreigners from Americans, so it pulled Fable 5 and Mythos 5 for everyone.
Let me walk you through what happened & explain what this means for us.
This week Anthropic launched two new models, Fable 5 and Mythos 5.
Both sit on the same very capable base. Mythos is the heavy one. It is so good at finding software security holes that Anthropic kept it limited to a small set of trusted partners under a program called Glasswing.
Mozilla said it fixed hundreds of security bugs using an early version of it. That cyber skill is exactly what makes the model valuable, and exactly what makes a government nervous.
On Friday, US Commerce Secretary Howard Lutnick sent a letter to Anthropic's CEO putting both models under export control.
By one account, the trigger was another company claiming it had jailbroken Mythos. The government had also tried earlier to get Anthropic to delay the launch and failed. The letter came right after that.
Anthropic complied within hours but pushed back hard in public.
It says the so-called jailbreak is narrow. Other public models like GPT-5.5 already do this, and security defenders use it every single day.
Anthropic's argument is simple. If a minor finding like this can yank a model used by hundreds of millions of people, then no frontier model from anyone could ever stay online.
The company also built Fable with extra guardrails and a 30-day data retention rule specifically to catch and shut down misuse fast, so being recalled over this stings even more for them.
Now, what does this mean for you and me, building on these tools from Bengaluru or Pune.
Day to day, almost nothing changes right now. Fable 5 and Mythos 5 are gone for everyone, but Claude Opus 4.8, Sonnet and Haiku all still work. Your normal Claude usage is fine. ChatGPT and Gemini are fine.
But the deeper story is that the most powerful AI is now being treated like advanced chips. A controlled export.
The same way the US decides which Nvidia chips India is allowed to buy, it can now decide which AI models India is allowed to use.
And the word "foreign national" puts all of us in one bucket, friend of America or not.
Remember that Anthropic's own CEO was in New Delhi in February, talking up AI for India. That closeness did not stop the tap from being turned off.
So this can lead to interesting ramifications.
One. A two-tier AI world.
US persons get the top tier, the cyber-capable frontier. The rest of us get the model one notch below, and even that can vanish overnight. This order landed at 5:21 PM with zero warning and the model was gone the same evening.
Two. A massive reason for India to build its own.
Every minister and founder talking about sovereign AI just got handed proof. If a tool can be switched off from a foreign capital you do not vote in, you cannot put anything critical on top of it.
Not a bank, not a hospital system, not a government service, nothing.
Three. A lesson for builders.
Do not wire your product to a single model from a single country and assume it stays.
Keep a fallback model ready to swap in. Abstract your prompts so you are not married to one provider. Treat model access like a supply chain with risk.
Four. Slower, messier launches ahead.
If governments can pull a model over a narrow finding, every lab will get more cautious about what it ships and where.
Expect more launches that are geo-gated from day one, where India is simply not on the list at the start.
I use these tools all day and I am not about to stop. I am not panicking and neither should you. None of this breaks your workflow tonight.
But if you are building on American AI from India, treat today as the day to plan for access being pausable.
Keep a backup. Read the sovereign AI conversation a lot more seriously than you did last week.
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…