Anthropic is bringing its two most capable Claude models back online. On June 30, 2026, the company announced that the Trump administration had lifted the US export controls on Claude Fable 5 and Mythos 5, ending an 18-day blackout that began days after the models first shipped. Fable 5, Anthropic's most capable widely released model, is returning globally starting July 1, while Mythos 5 comes back only to approved US organizations. For teams that built reasoning and agentic workflows around Fable 5 and then watched it vanish mid-project, the wait is over, but the model comes back with new identity checks, credit-based pricing, and a fresh safety filter bolted on.
What Anthropic Restored
The June 30 decision reverses a Commerce Department order that had forced Anthropic to cut off both models worldwide. As confirmed across outlets, the government removed the restrictions after Anthropic addressed the security concern that triggered them. The comeback is staged rather than instant: regulators had already cleared Mythos 5 on June 27 for a limited set of trusted partners, and the June 30 action widens that into a full restoration of Fable 5 for general use. Anthropic says access returns in phases, so not every account lights up at the same moment.
The 18-Day Blackout: A Timeline
The whole episode compressed a launch, a shutdown, and a restoration into under three weeks. Here is how it unfolded, per Anthropic and contemporaneous reporting.
| Date | What happened |
|---|---|
| June 9, 2026 | Fable 5 and Mythos 5 launch publicly with safety guardrails. |
| June 12, 2026 | Commerce order bars access for any foreign national, effective immediately. Unable to verify nationality in real time, Anthropic shuts both models down for everyone. |
| June 27, 2026 | Commerce clearance letter allows limited Mythos 5 redeployment to trusted partners. |
| June 30, 2026 | Export controls on both models lifted. Anthropic publishes its restoration plan and a new safety classifier. |
| July 1, 2026 | Fable 5 begins returning globally; Mythos 5 returns only to approved US organizations. |
We covered the shutdown as it happened in Anthropic Disables Fable 5 and Mythos 5 on US Order, and the original launch, including how Fable 5 stacks up against Opus 4.8, in Claude Fable 5 vs Opus 4.8: What Creators Get.
Fable 5 vs Mythos 5: What Each Model Does
The two models share the same underlying capabilities but return on different terms. Fable 5 is the generally available model with safety classifiers; Mythos 5 is the same capability without them, for approved partners only.
| Fable 5 | Mythos 5 | |
|---|---|---|
| Best for | Most demanding reasoning and long-horizon agentic work | Same capability as Fable 5, without safety classifiers |
| Who it is for | Developers and teams building agents | Approved US organizations via Project Glasswing |
| Return scope | Global, starting July 1 | US approved partners only, staged review |
| Access model | Identity verification plus paid usage credits | Government-reviewed partner allowlist |
Mythos 5 previously reached roughly 200 firms through Project Glasswing, including Apple, Google, Cisco, Nvidia, and Microsoft, per Fortune. For everyone outside that circle, Fable 5 is the practical return.
The New Access Rules: KYC, Credits, and Geography
Fable 5 does not simply switch back on inside your existing plan. Code found in Anthropic's iOS app points to identity verification as a gate before access is granted, and app strings pointed to usage-credit billing. In practice, Anthropic is including Fable 5 for up to 50% of weekly usage limits on Pro, Max, Team, and select Enterprise plans through July 7, then moving to metered usage credits at $10 per million input tokens and $50 per million output.
Three practical consequences follow. First, expect a know-your-customer step, so have identity documents ready. Second, budget for metered credit usage instead of a flat monthly allotment, which changes the cost math for heavy Fable 5 pipelines. Third, the verification layer looks designed to enforce geography, so users outside the US should confirm regional availability before assuming the model is live for their account.
How to Bring Your Fable 5 Workflow Back Online
If Fable 5 sat in your production pipeline, resist flipping everything back on at once. Work through it in order.
- Confirm the model is actually responding for your account. Access returns in phases, so check your plan's model list and send a small test request before trusting it with production work.
- Complete identity verification early. If you are prompted for KYC, finish it now so credits are ready when you need them, rather than mid-deadline.
- Reload credits and set a budget. Because Fable 5 runs on paid usage credits, estimate your per-project token spend and top up before a large job, not during.
- Verify regional access if you are outside the US. Confirm availability in your region before rebuilding around the model, since the rollout and the verification layer both enforce geography.
- Keep a fallback model wired in. Leave a secondary model configured until Fable 5 has been stable for your account for a few days, in case the phased rollout reaches you late.
- Re-open paused projects last. Only resume in-progress Fable 5 work once the steps above check out, so you do not stall a second time.
Why the Controls Happened
The shutdown was not arbitrary. The trigger, according to reporting, was a jailbreak: Amazon researchers found a way to bypass Fable 5's safeguards and prompt it to identify software vulnerabilities, in one case producing code showing how a flaw could be exploited. The June 12 order told Anthropic to cut off both models for any foreign national, and because the company could not check every user's nationality in real time, it pulled the models entirely.
To reopen, Anthropic trained a new safety filter, a classifier that watches for that exact technique and blocks it, which the company says now stops the method in more than 99% of attempts as of its June 30 write-up. Anthropic also said it is deepening collaboration with the US government, including giving designated agencies early access to frontier models and their safeguards before public release, sharing threat intelligence, and working toward a common security standard across AI developers. The restoration is as much a policy milestone as a product one.
Frequently Asked Questions
Is Fable 5 available again?
Yes. Anthropic began restoring Fable 5 globally on July 1, 2026, after the US lifted export controls on June 30. Access returns in phases, so it may not be live for every account at the same time.
Do I need to verify my identity to use Fable 5?
Likely yes. Code in Anthropic's iOS app indicates identity verification (a KYC step) is required before Fable 5 access is granted. An app string states that credits are added once identity is verified.
Is Fable 5 included in my Pro or Max plan?
Through July 7, 2026, Fable 5 is included for up to 50% of weekly usage limits on Pro, Max, Team, and select Enterprise plans. After that, access moves to metered usage credits at $10 per million input tokens and $50 per million output.
What is the difference between Fable 5 and Mythos 5?
Fable 5 is Anthropic's most capable widely released model, built for demanding reasoning and long-horizon agentic work, and it is returning globally. Mythos 5 shares the same capabilities without the safety classifiers and returns only to approved US organizations through Project Glasswing.
Why were the models shut down in the first place?
A jailbreak discovered by Amazon researchers let Fable 5 bypass its safeguards and surface software exploit details. A June 12 Commerce order barred foreign-national access, and Anthropic shut both models down because it could not verify nationality in real time.
Are users outside the US able to use Fable 5?
Confirm before you rebuild. Fable 5 is returning globally, but the identity-verification layer appears designed to enforce geographic limits, so regional availability should be checked per account.
