How do I access Claude Fable 5?
Two confirmed routes, one phased subscription rollout, and one correction the launch summaries keep getting wrong.
Claude Fable 5 is reachable through the Claude API under the model id claude-fable-5, fully available since the June 9, 2026 launch with a one-million-token context window, and through Anthropic's claude.ai apps. On Pro, Max, Team, and Enterprise the rollout is phased: free June 9 to 22, then usage credits. Anthropic did not announce Amazon Bedrock, Google Vertex AI, or Microsoft Copilot availability for Fable 5 at launch, so treat those as unconfirmed, and Mythos 5 is not publicly accessible. Access is the easy part; the structured First Brain you bring decides what the model is worth.
You access Claude Fable 5 in two ways today: through the Claude API, where it has been fully available since launch under the model id claude-fable-5, and through Anthropic’s own apps on claude.ai. On the paid subscription plans, the rollout is staged, so what you get depends on when you are reading this. Anthropic has confirmed the API and its apps; it has not, as of launch, confirmed availability on Amazon Bedrock, Google Vertex AI, or Microsoft Copilot, so treat those as unannounced rather than assumed. The restricted sibling model, Claude Mythos 5, is not publicly accessible at all. Here is each route, who it suits, and the one detail most launch summaries get wrong.
The short answer
Developers reach Fable 5 through the API immediately; everyone else reaches it through claude.ai, with paid plans on a phased schedule. Anthropic described the model as fully available through the Claude API from launch day, alongside access in its consumer apps. That is the dependable part. The part that shifts over time is how the subscription tiers treat it, which Anthropic laid out as a deliberate, dated rollout rather than a flat “it is included.”
Through the Claude API
This is the most direct and flexible route. If you build with Anthropic’s API, you call Fable 5 exactly as you would any other model, by passing the model id claude-fable-5, and you get the full one-million-token context window at standard pricing. Pricing is usage-based at 10 dollars per million input tokens and 50 per million output, with a 50 percent discount available through the Batch API for work that is not time-sensitive.
The API is the right route when you are building a product, running batch jobs, automating a workflow, or wiring the model into an agent that needs to act over many steps, which is where Fable 5’s longer autonomous runs matter most, a pattern explored in agents and the delegation of thought. It is pay-as-you-go, so there is no subscription to worry about; you are billed for the tokens you use. For most developers, the API is simply where Fable 5 lives.
Getting started is the same as with any Claude model: an API key, the official SDK or a plain HTTP call, and the model id in the request. The one practical wrinkle at launch is rate limits, since demand for a new frontier model is high and your tier’s per-minute limits may bind sooner than usual. If you hit them, the standard responses apply: retry with backoff, batch non-urgent work, or fall back to Opus 4.8 for the parts of the job that do not need the frontier.
On claude.ai and the paid plans
For non-developers, Fable 5 appears in Anthropic’s apps, and the subscription story is where timing matters. Anthropic announced a phased approach for the Pro, Max, Team, and Enterprise plans: for a launch window running June 9 to 22, 2026, Fable 5 was included at no extra cost. From June 23, it was removed from the standard plan allotment and made available through usage credits instead, with Anthropic stating it aims to restore Fable 5 as a standard feature later, capacity permitting.
The reason for the staging is capacity, not marketing. A model this much in demand at launch strains supply, and metering it through credits after the free window is how Anthropic keeps it available without degrading service. The practical takeaway is to check your plan’s current terms rather than assume, because this is exactly the kind of detail that changes month to month. If you are on a paid plan and Fable 5 is not showing as included, usage credits are the expected path.
| Access route | Available at launch | Best for | Notes |
|---|---|---|---|
Claude API (claude-fable-5) | Yes | Developers, products, agents, batch jobs | Pay-as-you-go; 1M context; Batch API 50% off |
| claude.ai apps | Yes | Individuals, writing, research, chat | The non-developer route |
| Pro / Max / Team / Enterprise | Phased | Subscribers | Free June 9-22, then usage credits |
| Amazon Bedrock / Google Vertex | Not stated for Fable 5 | Cloud-platform teams | Claude models broadly are on both; Fable 5 unconfirmed |
| Microsoft Copilot | Not stated | Microsoft-stack users | No announced availability at launch |
What about Bedrock, Vertex AI, and Copilot?
Here is the correction worth making, because it circulates the other way. At launch, Anthropic confirmed Fable 5 on its own API and apps but did not announce availability on Amazon Bedrock, Google Vertex AI, or Microsoft Copilot. Claude models broadly are offered on Bedrock and Vertex, which Anthropic’s own pricing documentation covers, so it is reasonable to expect Fable 5 to reach those platforms, but expecting and confirmed are different things. Until Anthropic or the platform says so, treat partner-cloud access to Fable 5 specifically as unannounced. If your stack depends on Bedrock or Vertex, plan against Opus 4.8 or another confirmed model there and watch for the Fable 5 listing rather than building on an assumption.
What about Mythos 5? It is not public
The most common access confusion is thinking Mythos 5 is a buyable upgrade. It is not. Mythos 5 is the same underlying model as Fable 5 with some safeguards lifted, and Anthropic restricts it to vetted partners, deploying it initially through a trusted-access program called Project Glasswing aimed at cyber defense and authorized research. There is no consumer or general-developer route to Mythos 5, and that is by design. For essentially everyone, the model you can access is Fable 5, and the difference between the two is worth understanding before you go looking for access you cannot get. The full distinction is covered in Fable 5 versus Mythos 5.
Which route is right for you?
Match the route to what you are doing. If you write code or build products, use the API; it is the most capable, most controllable, and the only route that gives you batch discounts and agent integration. If you mainly think, write, and research through a chat interface, use claude.ai and watch your plan’s terms around the credit rollout. If you run on a partner cloud, wait for a confirmed Bedrock or Vertex listing before committing. And if you believed you needed Mythos 5, you almost certainly do not; Fable 5 is the model behind the headlines.
Whichever route you take, the access question is the easy half. The harder and more durable question is what you do with the model once you have it, and that depends far less on the channel than on the structure of the mind using it.
How access changes nothing about the real bottleneck
A frontier model on tap does not, by itself, make anyone a better thinker; it makes them a faster one, in whatever direction they were already pointed. The bottleneck moves from “can I reach the model” to “do I know what to ask it, can I judge its answer, and do I hold enough context to direct it well.” Those are not access problems. They are structure problems, and they are solved by a connected internal model of what you know, a biological knowledge graph dense enough to frame sharp questions and catch confident errors.
This is First Brain before Second Brain applied to the moment a powerful tool becomes easy to reach. The people who get the most from Fable 5 are not the ones with the best access; they are the ones with the most structure, the same reason cognitive augmentation starts with your own biology rather than with the tool. Securing access is worth doing. Building the First Brain that makes the access pay off is the part that actually changes your results, and the method for building it is the core of Building Your First Brain, free for the first 1,000 readers.
Key takeaways: accessing Claude Fable 5
Claude Fable 5 is reachable today through the Claude API under the model id claude-fable-5, where it has been fully available since the June 9, 2026 launch with the full one-million-token context window, and through Anthropic’s claude.ai apps. On the Pro, Max, Team, and Enterprise plans the rollout is phased: included free from June 9 to 22, then available through usage credits, with Anthropic aiming to restore standard inclusion later. Anthropic did not announce Amazon Bedrock, Google Vertex AI, or Microsoft Copilot availability for Fable 5 at launch, so treat those as unconfirmed, and Claude Mythos 5 is not publicly accessible at all. The honest framing: access is the easy part, and the structured First Brain you bring decides what the model is actually worth to you.
Frequently asked questions
How do I access Claude Fable 5?
Two ways are confirmed: the Claude API, where it is fully available under the model id claude-fable-5 with a one-million-token context window, and Anthropic’s claude.ai apps. On paid plans the rollout is phased, free at launch then moving to usage credits, so check your plan’s current terms. Anthropic did not announce Bedrock, Vertex, or Copilot availability for Fable 5 at launch. The restricted Mythos 5 is not publicly accessible. For developers the API is the natural home; for everyone else, claude.ai. Getting the most from it, though, depends on the structured First Brain you bring to it.
Is Claude Fable 5 available on Amazon Bedrock or Google Vertex AI?
Not as an announced feature at launch. Anthropic confirmed Fable 5 on its own API and apps but did not state availability on Amazon Bedrock, Google Vertex AI, or Microsoft Copilot. Claude models broadly are offered on Bedrock and Vertex, so Fable 5 may follow, but until Anthropic or the platform confirms it, treat partner-cloud access as unconfirmed. If your stack depends on those platforms, plan against a confirmed model like Opus 4.8 there and watch for a Fable 5 listing rather than assuming it is present.
Is Claude Fable 5 free on Claude Pro or Max?
It was, for a launch window. Anthropic included Fable 5 at no extra cost on Pro, Max, Team, and Enterprise from June 9 to 22, 2026, then moved it to usage-credit access on those plans from June 23, with a stated aim to restore standard inclusion later as capacity allows. Because this is a dated, capacity-driven rollout, the honest answer depends on when you ask, so check your plan’s current terms rather than relying on the launch-week arrangement.
Do I need Claude Mythos 5, and how do I get it?
Almost certainly not, and you generally cannot get it. Mythos 5 is the same underlying model as Fable 5 with some safeguards lifted, restricted to vetted partners through a trusted-access program for cyber defense and authorized research. There is no public or general-developer route to it. For nearly every use, Fable 5 is the model in question, and it is fully accessible. If you thought you needed Mythos 5, the difference between the two is worth reading before assuming a restricted model is what your work requires.
What is the cheapest way to use Claude Fable 5?
For variable or batch workloads, the Batch API, which applies a 50 percent discount to both input and output tokens, is the cheapest route on the API. Prompt caching cuts the cost of repeated context further. For individuals, a paid plan can be cheaper than metered API use within its included limits, though the launch credit rollout complicates that comparison. The most underrated cost lever, though, is precision: a clear, well-structured request from someone who knows what they want spends far fewer tokens than a vague one, which is a First Brain skill before it is a billing one.
Dive deeper in
What is Claude Fable 5?
Claude Fable 5 is Anthropic's newest, most capable model (June 2026): state-of-the-art on most benchmarks, 1M context, $10/$50, model id claude-fable-5.
How much does Claude Fable 5 cost?
Claude Fable 5 is $10 per million input tokens and $50 output, double Opus 4.8. 1M context at standard rates; Batch API halves it; subscriptions are phased.
What is the difference between Claude Fable 5 and Claude Mythos 5?
Fable 5 and Mythos 5 are the same model governed two ways: Fable is safeguarded and public; Mythos has safeguards lifted and is restricted to vetted partners.