Claude Fable 5 is here. What it means for your work, in plain terms.

Update, June 28, 2026. Since this post went live on June 9, Claude Fable 5 was taken offline worldwide. On June 12, the US Commerce Department issued an export-control directive and Anthropic suspended Fable 5 globally, making this the first time the US has applied export controls to a live commercial AI API. As of today, the model is still serving zero traffic. Mythos 5, the restricted sibling, was partially restored on June 27 for roughly 100 vetted US critical-infrastructure organizations only. Fable 5 itself has not been restored anywhere. The most likely return path is US-citizens-first, with EU and other international users expected to wait significantly longer with no firm date. Anthropic is pointing everyone to Opus 4.8 in the meantime. The rest of this post covers what Fable 5 is capable of and what to do while it is down.

---

Anthropic released Claude Fable 5 on June 9. It is the most capable model they have ever made publicly available. Access was suspended globally on June 12 before most people had a chance to test it in a real workflow. This post covers what Fable 5 can do, what the suspension means, and what to do right now.

What is Fable 5, in one paragraph

Anthropic built a model called Mythos 5 that they consider too capable to release openly, keeping it restricted to vetted security and biomedical researchers. Fable 5 is the same model with a safety layer on top. Requests touching dangerous territory (think bioweapons, cyberattack tooling) get quietly rerouted to the older Opus 4.8 model. Anthropic says this happens in fewer than 5% of sessions. For consulting, finance, sales, and marketing work, you would essentially never hit it. The export-control directive that caused the June 12 suspension was triggered specifically by the model's ability to find software vulnerabilities in code. Anthropic has publicly noted that the same capability exists in other deployed models.

What it is best at

The pattern across every early benchmark and test: Fable 5 pulls ahead on long, complex, multi-step work. On short, simple prompts, the difference from Opus 4.8 is hard to feel.

Where the gains were clear and are still what to expect when access returns:

Financial analysis. Fable 5 scored highest of any model on Hebbia's Finance Benchmark, which tests reasoning at senior-analyst level: reading documents, interpreting charts and tables, solving problems across them. Trading firm IMC reported it passed their trading-analysis evaluations almost across the board.

Data analysis. First model to break 90% on Hex's analytics benchmark, a 10-point jump over Opus 4.8 on long-running analytical tasks that need judgment, not just computation.

Large document sets. It holds focus across millions of tokens and writes its own working notes as it goes. If your work involves a 40-page report, six competitor documents, and a regulatory framework all at once, this is the upgrade you will feel.

Reading visuals. A meaningful jump in chart, table, and screenshot interpretation. In one test it rebuilt a working web app from screenshots alone. Practical version: it reads the graphs in your client's PDF correctly more often.

Contract review. Lawyers in blind tests found its contract redlines matched or beat specialized legal AI tools. It is already integrated inside Harvey, the legal AI platform.

Long-running tasks. Anthropic markets "days-long" autonomous work. One caveat the launch coverage often skipped: that applies to automated pipelines built on the API, not to chatting in the Claude app. In normal use, what you get is a model that stays coherent much deeper into a long working session before it starts losing the thread.

What it is not better at

Quick drafts and short prompts. Early reviewers at Every.to found the model's slower, deeper thinking "works against the rapid-iteration rhythm most writing depends on". If your main use is drafting emails and punching up copy, Opus 4.8 is faster and costs about half as much.

Marketing and sales writing specifically. No independent test of proposal writing, marketing briefs, or sales decks surfaced before the suspension. The finance and analytics gains were confirmed. The writing gains were not.

Finding a use for it at all. The same Every.to review noted that heavy AI users found it paradigm-shifting on their hardest tasks, while lighter users "struggled to find something to use it for." If you have not yet built AI into a real workflow, a stronger model does not change that. The bottleneck is the workflow, not the model.

The data-residency picture for EU users

When Fable 5 does return, there is a GDPR consideration worth knowing now so you can plan.

On the direct Anthropic API, Mythos-class models carry mandatory 30-day data retention with no EU data residency option. That was true when the model launched and will likely remain true on return. The compliant EU routes are:

AWS Bedrock, EU Stockholm region. This was live before the suspension and is the one to watch. It provides EU data residency and works the same way as the US API for non-regulated prompts. Source: AWS blog on Fable 5 availability.

Claude Enterprise with EU data storage. Anthropic's Enterprise tier offers data hosted in Frankfurt and Dublin. If you are on Enterprise, confirm the EU-region option with your account contact before you start routing client data through any Mythos-class model.

If you are building workflows now under GDPR constraints, set them up to route through Bedrock or Enterprise from the start. Switching later is friction you do not need.

Is it available, and what will it cost

Right now: no. Fable 5 is suspended worldwide as of June 12, 2026, by a US Commerce Department export-control directive. Selecting it in a model picker returns "Claude Fable 5 is currently unavailable." Use Opus 4.8 (model id: claude-opus-4-8).

The June 9-22 free window has passed. Paid plan subscribers (Pro, Max, Team, Enterprise) had free access during that window. It closed before most people could use it.

When it returns, the API pricing was: $10 per million input tokens and $50 per million output tokens, roughly 2x Opus 4.8. Pricing on flat subscription plans (Pro, Team, Enterprise) has not been announced for the post-suspension return.

Who gets access first. Anthropic's new identity-verification policy (government ID required) takes effect July 8, widely read as the mechanism for a US-citizens-first restoration. EU, Canadian, Australian, and Indian users have no exemption in place and should plan on waiting longer. An August 1 executive-order deadline for a "covered frontier model" framework is the main policy milestone that could open broader access. Prediction markets as of late June priced roughly 68-71% odds of restoration before July 1, but treat that as speculation, not a date.

What to do while it is down

Use Opus 4.8. For most day-to-day work, including financial analysis, document summarization, writing, and research, Opus 4.8 is capable and available today. The gap between Opus 4.8 and Fable 5 was most visible on the hardest multi-document analytical tasks. Everything else, Opus 4.8 handles well.

If EU data residency matters to you, set up the right route now. AWS Bedrock with the Stockholm region endpoint, or Claude Enterprise with EU hosting, are the options to establish before Fable 5 returns. Do not wait until return day to figure out your compliant path.

Build your workflows to be model-agnostic. The Fable 5 suspension is the clearest possible signal that building a workflow tied to a single model ID is a fragile choice. If you are setting up templates, automations, or client-facing AI pipelines, build in an abstraction layer so you can swap models without rebuilding. That is true for Anthropic models, and for any other provider.

Line up your hardest task for when it returns. If financial modeling across multiple documents, large contract sets, or multi-source research is where you feel the limits of current models, Fable 5 is the one to test when it comes back. Get clear on what that task is now so you can evaluate it properly on day one of restored access.

Fable 5 will likely return, though when and for whom first is not settled. When it does come back, it will probably be US users first, international users later, with EU access among the last to be restored. There is no firm date from Anthropic. Check anthropic.com/news for updates.

---

Want to build AI into your actual workflows, your templates, your client work, your data? That is what we do in our hands-on workshops. You bring a real workflow, you leave with it running. Reserve a spot at aibl.to.