How to Access Claude Fable 5: API, Claude Apps, and Plans

Anthropic shipped Claude Fable 5 on June 9, 2026, and the first question every developer and subscriber is asking is simple: how do you actually get your hands on it? The good news is that there is more than one path to access Claude Fable 5 right now, and at least one of them is free for a limited window. This guide walks through every route, who each one is for, what it costs, and when the rules change. If you have used Anthropic’s recent releases before, the setup will feel familiar, much like getting started with the Claude Opus 4.8 API.

Claude Fable 5 is the new public flagship. There is also a sibling model, Claude Mythos 5, that most people cannot get, and the difference matters, so this article covers both. By the end you will know exactly which door to walk through and what to expect on the other side.

TL;DR

Claude Fable 5 is available today through the Claude API with the model ID claude-fable-5. From June 9 to June 22, 2026, it is also included free on Pro, Max, Team, and seat-based Enterprise plans. On June 23, 2026, it leaves those plans and requires usage credits. Claude Mythos 5 stays restricted with no public sign-up.

The two ways to access Claude Fable 5

There are two practical routes to access Claude Fable 5 today. The first is the Claude API, which works for any developer with an Anthropic account. The second is your existing Claude subscription, which includes Fable 5 inside the apps you already use, but only during a short free window. The table below lays out the difference so you can pick the right one.

If you write code, the API is the route you want, because it gives you the model ID directly and works the same way as the rest of the Messages API. If you mostly use Claude through the web app, desktop app, or mobile, your subscription already has Fable 5 during the free window. A lot of teams use both: the apps for ad-hoc work, the API for anything that needs to run on a schedule or inside a product.

For a deeper look at what the model actually is and how it compares to its predecessors, see our companion piece on what Claude Fable 5 is. This article stays focused on access.

Access Claude Fable 5 through the Claude API

The API is the most direct way to use Fable 5, and it is available today. Three things stand between you and your first response: an account, an API key, and the right model ID. Here is the short version.

Step 1: Get an Anthropic API key

Head to the Anthropic Console and sign in or create an account. Once you are in, open the API keys section, create a new key, and copy it somewhere safe. You only see the full key once, so store it in a password manager or your secrets vault right away. If you plan to bill against an organization, make sure you are creating the key under the correct workspace, not your personal default.

A quick note on billing: the API is consumption based. You pay for what you send and receive. Fable 5 costs $10 per million input tokens and $50 per million output tokens, which is less than half the price of the earlier Claude Mythos Preview. If you are mapping out spend, our Claude Fable 5 pricing breakdown goes deeper than this access guide does. Pricing veterans will recognize the structure from the Claude Opus 4.8 pricing model.

Step 2: Set the ANTHROPIC_API_KEY environment variable

Anthropic’s tooling looks for your key in an environment variable named ANTHROPIC_API_KEY. Setting it once keeps the key out of your code and out of version control. On macOS or Linux:

export ANTHROPIC_API_KEY="your-key-here"

On Windows PowerShell:

$env:ANTHROPIC_API_KEY = "your-key-here"

For anything beyond a quick test, put the key in a .env file or a managed secret store instead of typing it into your shell each session.

Step 3: Send your first request to claude-fable-5

The model ID is exactly claude-fable-5. You send requests to the standard Messages API endpoint at https://api.anthropic.com/v1/messages, with two headers that Anthropic requires on every call: x-api-key for your key and anthropic-version set to 2023-06-01. Here is a minimal request you can paste into a terminal:

curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model":"claude-fable-5","max_tokens":1024,"messages":[{"role":"user","content":"Hello, Fable 5"}]}'

If your key is valid and your account has access, you get back a JSON response with the model’s reply. If you get a 401, the key is wrong or unset. If you get a 404 on the model, double-check the spelling of claude-fable-5. You can confirm the current model ID and capabilities against the official Anthropic models overview at any time.

That is the whole access flow for the API. This guide keeps the code minimal on purpose. For streaming, system prompts, tool use, and production patterns, read our full Claude Fable 5 API how-to, which picks up where this section leaves off.

Access Claude Fable 5 on Pro, Max, Team, and Enterprise plans

If you subscribe to Claude, you may already have Fable 5 without touching the API. From June 9 through June 22, 2026, Anthropic is including Claude Fable 5 at no extra cost on Pro, Max, Team, and seat-based Enterprise plans. During this window, you select it from the model picker in the Claude apps the same way you would choose any other model, and it counts against your normal plan usage rather than a separate bill.

The dates matter, so read this part carefully. The free inclusion is a launch window, not a permanent feature of those plans. On June 23, 2026, Claude Fable 5 is removed from Pro, Max, Team, and seat-based Enterprise plans. After that date, using Fable 5 on one of those plans requires usage credits. In practice that means the model does not disappear from your account, but accessing it shifts from “included” to “pay per use” through credits rather than your flat subscription.

Anthropic has said it plans to restore some standard access on these plans when capacity allows. There is no firm date for that, so treat the credit requirement as the rule from June 23 onward and the free window as the exception. If your team relies on Fable 5 for daily work, the safe move is to plan for credits after June 22 and adjust if Anthropic announces broader access later.

A few things worth knowing about the plan route. First, the free window applies to seat-based Enterprise plans; consumption-based Enterprise plans get Fable 5 through usage the same way the API does, so the June 23 change does not apply to them in the same way. Second, this is separate from the API. An API key bills independently of your app subscription, so a developer can keep calling claude-fable-5 through the API even after the plan window closes, paying the standard token rates. If your goal is to keep using the model without managing credits in the app, the API is the more predictable path. For ways to stretch your usage without overspending, our guide on using Claude Fable 5 for free covers the practical options, and the same thinking that applies to using Claude Opus 4.8 for free carries over here.

Can you access Claude Mythos 5?

Short answer for most readers: no. Claude Mythos 5 is the restricted sibling to Fable 5, and it is not publicly available. It runs on the same underlying model as Fable 5 but with some safeguards lifted, which is exactly why access is locked down.

Right now, Mythos 5 is available only to a small group through Project Glasswing, a collaboration with the US government. That group is made up of cyberdefenders and infrastructure providers, and for them the cyber safeguards are lifted so they can do defensive security work that the public model would route away. Anthropic has said biology researchers will get access soon as well, with biology and chemistry safeguards removed for that cohort while cyber safeguards stay on. Existing Claude Mythos Preview users can upgrade to Mythos 5.

There is no public sign-up form for Mythos 5. You cannot buy your way in through a plan or an API key. Anthropic has described a broader trusted-access program that is planned but not yet open, so the honest takeaway is that almost everyone reading this will use Fable 5, not Mythos 5. If you have a legitimate need that fits the cyberdefense or research profile, the path is through Anthropic’s official channels, not a self-serve console. The full details are in Anthropic’s Claude Fable 5 and Mythos 5 announcement.

The practical implication for builders: design around Fable 5. It is the model you can actually access, and for the overwhelming majority of use cases it is the right one. Mythos 5 exists for a narrow set of trusted partners, and treating it as your target model would be planning around something you cannot get.

What gets routed away on Fable 5

One behavior can surprise people the first time they hit it, so it is worth calling out before you build. On Fable 5, certain sensitive queries are not answered by Fable 5 at all. Queries that touch cybersecurity, biology and chemistry, or model distillation topics are routed to Claude Opus 4.8 instead, where the appropriate safeguards apply.

This routing is deliberate, and it is rare. Anthropic reports the safeguards trigger in fewer than 5% of sessions, so most workloads never see it. But if you are building in one of those sensitive domains and notice that responses feel different, or that the model behind a particular answer is not the one you expected, the routing is the likely reason. The response still comes back through the same API call; you just may be talking to Opus 4.8 for that specific turn.

If your application is sensitive to which model produced a given response, log the model field that comes back in each API response so you can see when routing happened. For context on the model your sensitive queries land on, our overview of what Claude Opus 4.8 is explains its strengths. The short version: this is a safety feature, not a bug, and it affects a small slice of traffic.

Test your Claude Fable 5 access with Apidog

Before you wire Fable 5 into an application, it is worth confirming that your API key works and that your account actually has access to the model. The fastest way to do that without writing throwaway code is to send a real request through an API client and inspect the response. Apidog is built for exactly this kind of check.

Here is a clean way to verify your access:

1. Create a new HTTP request in Apidog pointed at https://api.anthropic.com/v1/messages using the POST method.
2. Add the two required headers: x-api-key set to your key, and anthropic-version set to 2023-06-01. Add content-type: application/json as well.
3. Paste a small JSON body with "model": "claude-fable-5", a max_tokens value, and a one-line user message.
4. Send the request and read the response. A clean 200 with a model reply confirms your key and access are good. A 401 means the key is wrong; a 404 on the model means a typo or a workspace that lacks access.

Because Anthropic supports streamed responses, you can flip on streaming in your request body and watch Apidog render the tokens as they arrive, which is a quick way to confirm streaming works before you build it into your product. Once the request succeeds, save it to a collection so your whole team can reuse the same verified call instead of rebuilding it from scratch. Store the key as an Apidog environment variable so it never gets pasted into a shared request body.

This is the same workflow many developers already use for other model providers; if you have ever followed our guide on testing the ChatGPT API with Apidog, the steps will feel familiar, only the endpoint and model ID change. To get set up, Download Apidog and import the request above, or build it in a couple of minutes. Apidog also lets you generate client code from a working request, so once your access check passes you can export the call straight into your language of choice.

FAQ

Is Claude Fable 5 available now?

Yes. Claude Fable 5 launched on June 9, 2026, and is available immediately through the Claude API with the model ID claude-fable-5. It is also included in the Claude apps for Pro, Max, Team, and seat-based Enterprise subscribers during the free window that runs from June 9 to June 22, 2026.

Do I need to pay to access Fable 5?

It depends on the route. Through the API, you pay per token: $10 per million input tokens and $50 per million output tokens. Through a Pro, Max, Team, or seat-based Enterprise plan, Fable 5 is free from June 9 to June 22, 2026, with no extra charge beyond your existing subscription. After June 22, the plan route requires usage credits.

How do I access Fable 5 after June 23?

On June 23, 2026, Fable 5 is removed from Pro, Max, Team, and seat-based Enterprise plans, and using it on those plans requires usage credits from then on. The API route is unaffected: you keep calling claude-fable-5 at the standard token rates with no plan window to worry about. Anthropic has said it plans to restore some standard plan access when capacity allows, but there is no date for that yet.

Can I access Claude Mythos 5?

Almost certainly not. Claude Mythos 5 is restricted and has no public sign-up. It is available only to a small group through Project Glasswing, a US government collaboration covering cyberdefenders and infrastructure providers, with biology researchers getting access soon. Existing Claude Mythos Preview users can upgrade. A broader trusted-access program is planned but not open, so most readers should build around Fable 5.

What is the Fable 5 API model name?

The model ID is exactly claude-fable-5. You pass it in the model field of your request to the Messages API at https://api.anthropic.com/v1/messages. If you get a 404 on the model, check the spelling and confirm your workspace has access.

Which plans include Fable 5?

During the June 9 to June 22, 2026 window, Fable 5 is included at no extra cost on Pro, Max, Team, and seat-based Enterprise plans. Consumption-based Enterprise plans access Fable 5 through usage, the same as the API. On June 23, the included access on the seat-based plans ends and shifts to usage credits.

What happens to sensitive queries on Fable 5?

Queries about cybersecurity, biology and chemistry, or model distillation are routed to Claude Opus 4.8, where the relevant safeguards apply. This triggers in fewer than 5% of sessions, so most workloads never see it, but it explains why a sensitive query might be answered by a different model than you expected.

Do I need a separate key for the API and the apps?

Yes. Your Claude app subscription and your Anthropic API key are billed and managed separately. A developer can call claude-fable-5 through the API regardless of what plan window applies in the apps, which is why the API is the more predictable long-term route for production work.

Pick the route that fits how you work. If you are building anything that runs in code, get a key from the Anthropic Console, set ANTHROPIC_API_KEY, and send your first request to claude-fable-5 today. If you mostly use the Claude apps, you already have Fable 5 through June 22; just plan for credits after that. Either way, verify your access with a single request in Apidog before you build on top of it, and you will save yourself the surprise of a failed call halfway through a project.