Claude Fable 5 Is Here: What It Is, How To Use It, And When To Skip It

TL;DR: On June 9, 2026, Anthropic released Claude Fable 5, its most powerful generally available model ever. It sits above Opus 4.8 on nearly every benchmark and Every called it “the best coding model in the world” after early testing. The catch: it costs double Opus, runs slower, and is token hungry. Fable 5 is the safe, public version of Mythos 5, the same underlying model with safeguards lifted for vetted security partners. Reach for Fable 5 on big, multi-step jobs you can plan and walk away from. For quick everyday work, Opus 4.8 at half the price is still your daily driver.

What is Claude Fable 5?

Claude Fable 5 (model id claude-fable-5) is Anthropic’s new flagship model, released June 9, 2026. It is generally available across the Claude API, claude.ai on Pro, Max, Team, and Enterprise plans, and the major clouds including AWS, Google Cloud, and Microsoft Foundry.

The short version: it is the most capable model Anthropic has ever made available to the public, and it beats Opus 4.8 on almost every published benchmark. On agentic coding it scores 80.3 percent on SWE-Bench Pro versus 69.2 percent for Opus 4.8. On knowledge work it scores 1932 on GDPval-AA versus 1890. It is built for long-horizon work, which means it can plan a task across stages, spin up its own sub-agents, run for hours, and at high effort check and validate its own output before it hands anything back.

That last part matters. Fable 5 is not designed to be fast. It is designed to be thorough on jobs that would normally take you hours or days.

How do you use Fable 5?

Switching to Fable 5 takes about two seconds, and you do not need to touch the API to get it.

Desktop or web app: Click the model dropdown at the top of the chat and select Fable 5.
Claude Code in the terminal: Type /model and pick Fable 5 from the list.

There is one quiet behavior worth knowing. Fable 5 ships with safety classifiers. When a query lands on a safeguarded topic such as cybersecurity, biology, or chemistry, Fable 5 routes that specific answer to Opus 4.8 instead. For the overwhelming majority of work, coding, analysis, content, research, and agentic workflows, that handoff never fires and you never notice it.

Where Fable 5 fits in the Anthropic ecosystem

Here is the part most coverage skips. Fable 5 and Mythos 5 are the same model.

Anthropic did something on June 9 it had not done before. It shipped a single frontier model as two separate products. Mythos 5 is the version with safeguards lifted in sensitive areas, and it is not for sale. Access is restricted to vetted Project Glasswing partners, the cyberdefenders and critical infrastructure providers Anthropic works with. Anthropic describes Mythos 5 as having the strongest cybersecurity capabilities of any model in the world, which is exactly why it is gated.

Fable 5 is that same underlying model with the guardrails on, made safe for general release. So when you use Fable 5, you are running Anthropic’s frontier model with a seatbelt. That framing also clears up a common mix-up: there is no separate “Mythos app” you can download. Mythos is the locked version. Fable is the one you actually get.

What do early testers say?

Because Fable 5 only dropped on June 9, almost nobody has had real time with it yet. The one substantial early review comes from Every, the media and software company run by Dan Shipper, which got access about a week ahead of launch.

Their verdict was blunt: Fable 5 is the best coding model in the world. On Every’s Senior Engineer benchmark, the hardest coding test they run, Fable 5 scored 91 out of 100. Opus 4.8 scored 63. GPT-5.5 scored 62. Every reported building projects that would normally take hours to days from a single one-shot prompt.

Treat that as one strong data point, not the final word. A model this new has not been stress-tested across thousands of real workflows yet. The honest read today is: the early signal is excellent, and the wider verdict is still coming.

Who should actually use Fable 5?

This is where the hype needs a filter. Fable 5 is a premium tier, not a free upgrade, and it is not the right pick for everyone.

Reach for Fable 5 if you:

Run long, multi-step agentic tasks that go for hours
Ship serious code or build full projects in one pass
Do deep research runs or complex planning work
Have a single big job that genuinely needs the strongest model

Stick with Opus 4.8 if you:

Ask quick questions throughout the day
Do everyday vibe coding on a normal setup
Make small edits and back-and-forth chat
Want a sensible default at half the price

If you are a knowledge worker or casual builder, you may not notice a huge difference on day-to-day tasks, and you will burn through credits faster for the privilege. There is no shame in keeping Opus 4.8 as your daily driver and saving Fable 5 for the one job a week that actually moves the needle.

The cost and the June 23 deadline

Fable 5 costs 10 dollars per million input tokens and 50 dollars per million output. That is double Opus 4.8’s 5 and 25. A 90 percent prompt-caching discount applies to input, which softens the hit when you reuse context, but the base reality is simple: Fable 5 is roughly twice the price and more token hungry, so a sloppy run gets expensive.

There is also a timeline to plan around. Fable 5 is included free on subscription plans from June 9 through 22. From June 23 it runs off usage credits, then returns to standard plan access once capacity allows. Worth being precise here, because some early coverage claimed Fable 5 goes “API only” on June 23. That is not correct. It moves to usage credits on plans, it does not disappear from the apps.

The practical move: use the free window to put Fable 5 on a real project so you know whether it earns its keep before credits start counting.

How to get the most out of Fable 5

Fable 5 rewards a good plan more than a clever prompt. Its strength is long-horizon execution, so the worst thing you can do is hand it a vague task and hope.

Write a clear, step-by-step plan or spec for the job first.
Point Fable 5 at the plan and let it run uninterrupted.
Use it for the meaty work that justifies the cost, not your daily chat.

Set it up well and you get a model that will work through a multi-day project, manage its own sub-tasks, and validate its output, all from a single kickoff.

FAQ

Is Fable 5 better than Opus 4.8? On benchmarks, yes, Fable 5 sits above Opus 4.8 on nearly every published test. In practice it is better at hard, long-horizon work and not meaningfully better at quick everyday tasks, where Opus 4.8 stays the smarter pick on cost.

Is Fable 5 the same as Mythos? Same underlying model, different product. Mythos 5 has safeguards lifted and is restricted to vetted security partners. Fable 5 is the safeguarded public version. You can use Fable. You cannot use Mythos.

How much does Fable 5 cost? 10 dollars per million input tokens and 50 per million output, double Opus 4.8, with a 90 percent prompt-caching discount on input. It is free on subscription plans from June 9 to 22, then runs on usage credits from June 23.

How do I switch to Fable 5? In the app, use the model dropdown and select Fable 5. In Claude Code, type /model and pick it.

Should I make Fable 5 my default model? For most people, no. Keep Opus 4.8 as your daily driver and switch to Fable 5 for big, multi-step jobs where the extra capability is worth the extra cost.

Want the simple decision rule on a single page? Grab the Fable 5 vs Opus cheat sheet and stop guessing which model to run.