Update — July 6, 2026: the included-in-subscription window ends tomorrow, July 7. From July 8, Fable 5 usage on Pro/Max/Team plans no longer counts toward weekly limits — it bills through prepaid usage credits at standard API rates ($10 / 1M input, $50 / 1M output). If you have Fable 5 wired into a coding agent on a subscription plan, set up credits now or pin a fallback model. Full breakdown: Fable 5 usage credits explained.
Anthropic shipped its most capable model ever on June 9, 2026, had it switched off by a U.S. export-control order three days later — and, after an 18-day standoff, brought it back on July 1. If you have been searching “is Claude Fable 5 back” or “is Fable 5 back yet,” the answer is now yes. Here is the clear, current picture: that it is live again, why it was suspended, what Anthropic changed to get it reinstated, where you can use it today, and what the whole episode means if you build on frontier models.
Want the full picture? Read our continuously-updated Claude Fable 5 launch guide — benchmarks, pricing, and everything Anthropic announced at release.
Is Claude Fable 5 back yet?
Yes. Claude Fable 5 — and its more tightly controlled sibling, Mythos 5 — returned on July 1, 2026. The turning point came a day earlier, on June 30, when the U.S. Commerce Department lifted the export-control directive it had imposed on June 12. That order had been in place for 18 days; with it withdrawn, Anthropic redeployed both models worldwide.
This is confirmed by Anthropic directly — see the company’s “Redeploying Fable 5” announcement — and corroborated by CNBC, The Hacker News, and others. If you still see “Fable 5 is back — June 18” claims on low-quality tracker sites, ignore them: the real return date is July 1, 2026.
What is Claude Fable 5?
Claude Fable 5 is the first generally-available model in Anthropic’s new top tier, which the company calls Mythos-class — a step above Claude Opus. Anthropic launched it on June 9, 2026 alongside Claude Mythos 5, describing Fable 5 as “the most capable model Anthropic has ever released publicly.”
The two models share the same underlying engine. Fable 5 is the version made safe for general use, with safeguards in place; Mythos 5 is the unrestricted version reserved for vetted, authorized users such as cybersecurity teams and infrastructure providers. Anthropic prices Fable 5 at $10 per million input tokens and $50 per million output tokens, with a 1M-token context window and up to 128K tokens of output. For the full launch details, see our Claude Fable 5 launch guide.
Why was Claude Fable 5 suspended?
On June 12, 2026 — just three days after launch — Anthropic received a directive from U.S. national-security authorities and disabled both Fable 5 and Mythos 5 globally. The order came from the Commerce Department’s Bureau of Industry and Security and invoked export-control powers, requiring Anthropic to obtain government permission before “exporting” the models to any foreign national anywhere in the world — including Anthropic’s own non-U.S. employees.
Because a consumer product serving hundreds of millions of people can’t reliably screen users by citizenship in real time, Anthropic switched the models off entirely rather than attempt selective enforcement. Every other Claude model stayed online.
The trigger was a jailbreak reported by researchers at Amazon. They found a way to bypass Fable 5’s safeguards by prompting it to read a codebase and identify software vulnerabilities — and in one case, the model produced code demonstrating how a vulnerability could be exploited. Anthropic characterized this as a “narrow, non-universal jailbreak” reproducible on other frontier models, and disagreed that a narrow jailbreak “should be cause for recalling a commercial model deployed to hundreds of millions of people” — but complied with the order while it worked to resolve the issue.
What Anthropic changed to bring Fable 5 back
The fix that ended the standoff was targeted rather than sweeping. Anthropic added defense-in-depth around the specific technique regulators flagged:
- A new safety classifier. Anthropic trained a classifier that detects and blocks the reported jailbreak technique in over 99% of cases. When a request trips the classifier, it is rerouted to Claude Opus 4.8 instead of being served by Fable 5 — so the flagged behavior can’t reach the more capable model.
- Independent government review. Researchers at the Commerce Department’s Center for AI Standards and Innovation (CAISI) tested both Anthropic’s prior and new safeguards before the export controls were lifted.
- A HackerOne channel. Anthropic opened a bug-bounty program on HackerOne where security researchers can responsibly report potential cyber jailbreaks, plus a wider safety margin and accepted government conditions.
Notably, the reroute-to-Opus-4.8 behavior mirrors Anthropic’s refusal-fallback pattern in the Claude API: when a safety classifier declines a Fable 5 request, developers can have it transparently re-served by a fallback model in the same call. If you build on Fable 5, it’s worth wiring that fallback in by default so a declined request degrades gracefully instead of failing outright.
Where you can use Claude Fable 5 now
Fable 5 is available again across Anthropic’s own surfaces — the Claude Platform (API), Claude.ai, Claude Code, and Claude Cowork — with cloud access on AWS, Google Cloud, and Microsoft Foundry being re-enabled. On the consumer and team plans (Pro, Max, Team, and select Enterprise), access is metered while capacity ramps back up:
| What | Detail |
|---|---|
| Return date | July 1, 2026 (export controls lifted June 30) |
| Surfaces | Claude Platform (API), Claude.ai, Claude Code, Cowork; AWS / Google Cloud / Microsoft Foundry re-enabling |
| Usage (Pro/Max/Team/Enterprise) | Counts toward up to 50% of weekly limits through July 7, then via usage credits |
| Pricing (API) | $10 / 1M input tokens, $50 / 1M output tokens |
If you moved a coding agent or pipeline off Fable 5 during the outage, you can point it back now — but keep the fallback you set up. The lowest-risk path is to leave Claude Opus 4.8 wired in as your fallback (it’s where the classifier reroutes flagged requests anyway) and re-test your prompts on Fable 5 before flipping production traffic back.
What the episode means for AI builders
Even with Fable 5 back, the three-week saga is being read as a precedent. It was one of the first times the U.S. government used export-control authority to switch off a commercially deployed AI model, and security and policy commentators were sharply divided on whether the bar applied here — a narrow, reproducible jailbreak — would, if generalized, stall future model launches across the industry.
The practical takeaway for anyone building on frontier models is unchanged, and if anything reinforced: don’t hard-code a single model into critical paths. Keep a tested fallback ready, abstract your model calls behind a thin interface, and assume any given model can become unavailable — for regulatory, capacity, or pricing reasons — on short notice. Teams that had done this rode out the Fable 5 outage by swapping to Opus 4.8; teams that hadn’t spent the downtime rewiring under pressure.
Building fast and short on engineers?
Codersera matches you with vetted remote developers who can build resilient, model-agnostic AI features — fast, and without the long hiring cycle.
Hire vetted remote developers →FAQ
Is Claude Fable 5 back online?
Yes. Anthropic redeployed Claude Fable 5 and Mythos 5 on July 1, 2026, after the U.S. Commerce Department lifted its export-control order on June 30. The models had been offline since June 12.
How long was Claude Fable 5 suspended?
The export-control order was in place for 18 days — from June 12 to June 30, 2026 — and Anthropic brought the models back online on July 1, an outage of just under three weeks.
Why did the U.S. government shut down Claude Fable 5?
A national-security export-control directive from the Commerce Department’s Bureau of Industry and Security required Anthropic to restrict the model from foreign nationals. Because it couldn’t filter users by citizenship at scale, Anthropic disabled Fable 5 and Mythos 5 worldwide. The order followed a jailbreak reported by Amazon researchers that got the model to identify software vulnerabilities and, in one case, produce exploit code.
What changed so Claude Fable 5 could return?
Anthropic trained a new safety classifier that blocks the flagged jailbreak technique in over 99% of cases and reroutes those requests to Claude Opus 4.8. Commerce’s Center for AI Standards and Innovation (CAISI) reviewed the safeguards, and Anthropic added a HackerOne channel for reporting cyber jailbreaks. Once the safeguards were validated, the export controls were lifted.
Where can I use Claude Fable 5?
It’s available on the Claude Platform (API), Claude.ai, Claude Code, and Cowork, with AWS, Google Cloud, and Microsoft Foundry access being re-enabled. On Pro, Max, Team, and select Enterprise plans it counts toward up to 50% of weekly usage limits through July 7, 2026, then moves to usage credits.
What is the difference between Claude Fable 5 and Claude Mythos 5?
They run on the same underlying model. Fable 5 is the safeguarded, generally-available version; Mythos 5 is the unrestricted version limited to vetted, authorized users. Both were suspended together and both returned on July 1, 2026.