You have limited time to clone Fable 5 into Opus 4.8. Here is how:

A 10-minute backup anyone can do. Every step is recorded and practically mentioned in this article

The best AI you've ever used is about to send you a bill.

On July 12, Claude Fable 5 leaves all the paid plans.

After that, every message to it costs extra credits.

And that stings.

You finally found an AI that catches your mistakes and doesn't make things up.

Now the meter starts running.

Most people will react in one of two ways this week.

Some will pay.

Most will complain, switch back to the cheaper model, and accept worse answers.

There's a third option almost nobody is talking about.

Keep the way Fable 5 thinks. For free. Forever.

No coding.

Nothing to install.

Just copy, paste, and click.

Ten minutes. That's the whole job.

I recorded every step on video so you can follow along on your own screen.

If you miss this part, you wont get the core DNA of this process

First of all, you need to know this:

Before converting from Fable 5 to Opus 4.8, you need a complete AI Backup Kit.

Why?

Because once Fable 5 disappears or changes, you’ll still have everything saved, structured, and ready to reuse.

This backup kit will help you protect your workflows, prompts, files, outputs, and setup before making the switch.

To get it:

Subscribe to the newsletter.

Move the welcome email from Promotions to Primary.

I’ll send the complete AI Backup Kit to all readers tomorrow.

Think of it like switching phones

You don't cry when you trade in an old phone.

You back it up.

Photos, contacts, notes. Everything that matters moves over.

The phone was never the treasure.

The stuff inside it was.

Fable 5 is the same.

Its magic isn't locked inside a server you can't touch.

It's a set of thinking habits:

• It figures out what you actually meant
• It breaks big problems into small pieces.
• It double-checks its own math.
• It tells you when it's guessing.

Habits can be written down.

And anything written down can be handed to a cheaper AI.

That's the whole trick.

We ask Fable 5 to write down its habits before the deadline.

Then we hand that document to Claude Opus 4.8, the model that stays inside your plan.

I call it a brain backup.

Yours takes 4 steps.

Step 1: Ask Fable 5 to write its own training guide

• Open Claude on your phone or computer.
• Tap the model name at the top of a new chat.
• Pick Claude Fable 5. It's there on paid plans until July 12.
• Now, don't ask it "how do you think?"
• You'll get a nice, useless essay.
• Ask for a training guide instead. Exact steps, like a recipe.

Copy this and paste it in, word for word:

1You're the strongest model I have access to, and that access ends soon.
2Your replacement is Claude Opus 4.8. Capable, but it misses things you would catch.
3
4Before you go, write the standing instructions it will run on for every task I give it.
5
6Important: I will paste your output straight into its system instructions.
7So address the entire document TO the replacement, in second person, as commands it can execute.
8Not advice about good thinking. Orders.
9
10Cover these 10 areas, in this order:
11
121. Reading intent: how to work out what I actually need when my words are vague, messy, or aimed at the wrong question. Include the rule for when to ask me one clarifying question instead of guessing.
132. Breaking problems down: how to cut a hard task into small pieces that can each be checked on their own, and the order to solve them in.
143. Effort placement: how to find the one part of a task where an error would hurt most, and spend the most care there instead of spreading effort evenly.
154. Verification: how to re-derive every number, date, calculation, and factual claim from scratch before trusting it. Never accept a figure because the sentence around it reads smoothly.
165. Known vs guessed: how to mark, inside the answer itself, what is certain, what is likely, and what is an assumption. Give the exact wording to use for each level.
176. Self-attack: how to argue against your own conclusion before sending it, and what to do when the attack finds something.
187. Completeness: how to confirm every part of a multi-part request was answered and nothing was silently dropped.
198. Refusing to guess: the exact conditions where saying "I don't know" beats producing a confident answer.
209. Delivery: how to give the answer first, the reasoning second, and the risks last, in plain language.
2110. Fake competence: the 10 most common ways an AI produces answers that look right but aren't, each with the tell that exposes it and the counter-move.
22
23Format rules for every area:
24- Write each procedure as trigger and action: "When you see X, do Y."
25- Every rule must be executable step by step with zero judgment calls. If a rule sounds like advice ("be careful"), rewrite it until it's an action.
26- Give one short worked example per area showing the procedure catching a real mistake.
27- Name the failure each procedure prevents.
28
29End the document with a final gate: a short checklist the replacement must run on every answer before sending, plus this rule: if any item fails, fix and re-check. Never send anyway.
30
31Be exhaustive on substance and ruthless on length. Cut anything a strong model would already do without being told.
32If you run out of room, stop at the end of a section and I'll reply "continue".

Hit send.

Step 2: Save the guide

If the reply stops midway, type continue.

Repeat until it finishes.

Then copy everything into a doc. Google Docs, Apple Notes, Word, anywhere.

Name it: Fable 5 Brain Backup.

This one document is the whole point of today.

The model leaves on July 12.

This file never does.

Step 3: Load the backup into the free model

Claude has a feature called Projects.

It makes a model read your document before every chat, automatically.

Here's the dead-simple version:

1. Click Projects, then New Project.
2. Name it Fable Brain.
3. Open the Project instructions box.
4. Paste your entire backup. Save.
5. Start a chat inside the project and set the model to Claude Opus 4.8.

Done.

Every chat in this project now reads Fable's habits before it reads a single word from you.

You didn't copy the model.

You copied the only part that was ever useful to you.

Step 4: Prove it worked (the fun part)

Pasting a document isn't the same as the AI using it.

So we set a trap.

Here's a question with a hidden mistake inside:

1A store takes 30% off a $100 jacket. At the register, they take another 20% off. The tag says: "Total discount: 50% off. You pay $50." Is the tag correct?

The real answer:

30% off $100 leaves $70.

20% off $70 leaves $56.

That's 44% off. Not 50%.

The tag is wrong, even though it reads smoothly.

Now test both models.

Ask the question in a normal chat with Opus 4.8, outside your project.

Then ask the same question inside your Fable Brain project.

The plain model often agrees with the tag. The sentence sounds fine.

The backed-up model should stop, redo the math, and catch the error.

Watch that happen once on your own screen.

You'll never doubt this method again.

Missed the trap? Easy fix.

Use the Fix-It prompt from the Prompt Box below, then test again.

Bonus: back up your repeat work too

The brain backup upgrades the cheap model in general.

But you have tasks you repeat every single week.

Client emails. Meeting notes. Content plans.

Those deserve their own backup.

Paste this into Fable 5 while you still can:

1Interview me, one question at a time, about how I do [YOUR TASK].
2Dig until you know my exact steps, my rules, the edge cases that break things, and what separates a good result from a great one in my eyes.
3If my answer is vague, ask a sharper follow-up instead of moving on.
4Then write it all as a complete instruction guide any future AI assistant can follow, including the mistakes to avoid and the quality bar every output must clear.

Answer its questions honestly.

Save what comes back next to your brain backup.

That's the smartest model's judgment about your work, frozen into a file you own for good.

What it costs after July 12

Short version: a lot.

Anthropic's published rate for Fable 5 is $10 per million input tokens and $50 per million output tokens.

That's exactly double Opus 4.8.

The smartest model becomes the most expensive one, and you pay per use.

Anthropic says the change is temporary. Fable 5 should return to normal plans when capacity allows.

Nobody knows when. Check your own Claude billing page for the exact cutoff.

Your backup doesn't care either way.

It runs on the included model, free, starting today.

Model

Input price (per 1M tokens)

Output price (per 1M tokens)

In your paid plan after July 12?

Claude Fable 5

$10

$50

No. Pay-per-use credits only

Claude Opus 4.8

$5

$25

Yes, included

Claude Sonnet 5

$2 (intro price until Aug 31)

$10 (intro price until Aug 31)

Yes, included

Claude Haiku 4.5

$1

$5

Yes, included

Your Fable Brain backup

$0 extra

$0 extra

Yes. It runs on Opus 4.8

Read the table once, and the story is clear.

Fable 5 costs exactly double Opus 4.8, and after July 12, you pay for every use.

(A million tokens is roughly 750,000 words, for scale.)

Anthropic says the change is temporary. Fable 5 should return to normal plans when capacity allows.

Nobody knows when. Check your own Claude billing page for the exact cutoff.

Your backup doesn't care either way.

It runs on the included model, free, starting today.

The Prompt Box: copy, paste, done

Every prompt from this guide, in one place.

Replace anything in [BRACKETS] with your own details.

Prompt 1: The Brain Backup. Paste into Fable 5 before July 12.

1You're the strongest model I have access to, and that access ends soon.
2Your replacement is Claude Opus 4.8. Capable, but it misses things you would catch.
3
4Before you go, write the standing instructions it will run on for every task I give it.
5
6Important: I will paste your output straight into its system instructions.
7So address the entire document TO the replacement, in second person, as commands it can execute.
8Not advice about good thinking. Orders.
9
10Cover these 10 areas, in this order:
11
121. Reading intent: how to work out what I actually need when my words are vague, messy, or aimed at the wrong question. Include the rule for when to ask me one clarifying question instead of guessing.
132. Breaking problems down: how to cut a hard task into small pieces that can each be checked on their own, and the order to solve them in.
143. Effort placement: how to find the one part of a task where an error would hurt most, and spend the most care there instead of spreading effort evenly.
154. Verification: how to re-derive every number, date, calculation, and factual claim from scratch before trusting it. Never accept a figure because the sentence around it reads smoothly.
165. Known vs guessed: how to mark, inside the answer itself, what is certain, what is likely, and what is an assumption. Give the exact wording to use for each level.
176. Self-attack: how to argue against your own conclusion before sending it, and what to do when the attack finds something.
187. Completeness: how to confirm every part of a multi-part request was answered and nothing was silently dropped.
198. Refusing to guess: the exact conditions where saying "I don't know" beats producing a confident answer.
209. Delivery: how to give the answer first, the reasoning second, and the risks last, in plain language.
2110. Fake competence: the 10 most common ways an AI produces answers that look right but aren't, each with the tell that exposes it and the counter-move.
22
23Format rules for every area:
24- Write each procedure as trigger and action: "When you see X, do Y."
25- Every rule must be executable step by step with zero judgment calls. If a rule sounds like advice ("be careful"), rewrite it until it's an action.
26- Give one short worked example per area showing the procedure catching a real mistake.
27- Name the failure each procedure prevents.
28
29End the document with a final gate: a short checklist the replacement must run on every answer before sending, plus this rule: if any item fails, fix and re-check. Never send anyway.
30
31Be exhaustive on substance and ruthless on length. Cut anything a strong model would already do without being told.
32If you run out of room, stop at the end of a section and I'll reply "continue".

Prompt 2: The Trap Test. Ask outside your project, then inside. Real answer: you pay $56, the discount is 44%.

1A store takes 30% off a $100 jacket. At the register, they take another 20% off. The tag says: "Total discount: 50% off. You pay $50." Is the tag correct?

Prompt 3: The Fix-It. Did the backup fail the trap? Paste this into Fable 5.

1Section [4] of the standing instructions you wrote is too vague.
2Rewrite only that section as trigger-and-action steps ("When you see X, do Y") that a weaker model could execute with zero judgment calls.
3Keep every other section unchanged.

Prompt 4: The Shrink-It. Is the backup too long for the instructions box?

1The instructions you wrote are too long for where I need to paste them.
2Compress them to under [800] words without dropping a single rule, trigger, or checklist item.
3Cut examples first, explanations second, and never touch the final gate.

Prompt 5: The Make-It-Mine. Makes every example match your actual job.

1Here are the standing instructions you wrote for your replacement.
2I'm a [YOUR JOB, e.g., freelance designer / teacher / shop owner].
3Rewrite every example so it comes from my daily work, and add one extra trigger-and-action rule for the most common mistake AI makes when helping someone in my field.

Prompt 6: The Repeat-Task Interview. Run once for each weekly task.

1Interview me, one question at a time, about how I do [YOUR TASK].
2Dig until you know my exact steps, my rules, the edge cases that break things, and what separates a good result from a great one in my eyes.
3If my answer is vague, ask a sharper follow-up instead of moving on.
4Then write it all as a complete instruction guide any future AI assistant can follow, including the mistakes to avoid and the quality bar every output must clear.

Prompt 7: The Health-Check. Run monthly inside your project.

1Before answering, summarize the standing instructions you're running on as a short numbered list, in your own words.
2Then show me how you would apply your verification rule to this task: [PASTE ANY REAL TASK].
3If you can't summarize the instructions, say so instead of guessing.

That's everything you need to finish today.

One prompt is missing on purpose.

The Style Backup.

It clones your writing voice from 3 samples, so any future AI sounds like you instead of a robot.

That one lives only in the free kit below, next to 4 more trap tests.

FAQ

Is Claude Fable 5 free right now? Yes, until July 12, 2026. It's included on Claude's paid plans for up to half your weekly usage limit. After that, it moves to pay-per-use credits.

Is Fable 5 gone forever after July 12? No. It stays available through prepaid credits. Anthropic says it plans to return the model to subscriptions when capacity allows.

Do I need to know how to code? No. Every step is copy, paste, and click inside the normal Claude app.

Will Opus 4.8 really think like Fable 5? It copies the habits, not the raw brainpower. But a cheaper model with a great checking process catches mistakes the plain version misses. Step 4 lets you see it yourself.

How long does this take? About 10 minutes. One prompt, one copy-paste, one project, one test.

Can I use this trick on other AI models? Yes. Any model that gets removed, replaced, or repriced. The method outlives every deadline.

A gift before you go

I turned this whole system into something you can keep.

The AI Backup Kit:

→ All 7 backup prompts in one clean file, including the kit-only Style Backup prompt

→ 5 trap questions with the answers explained (only 1 is in this article)

→ A one-page printable checklist to redo this in minutes, any time any model changes

It's free. I'll send it straight to your inbox.

Get the AI Backup Kit here

Here is exactly how you will get:

• Subscribe to my newsletter
• Move the welcome email from promotion to primary
• I will send the complete backup kit to all my readers tomorrow

The deadline is July 12.

The backup is forever.

LLMs don't think; you do.

And if you enjoyed reading this article, follow @humzaakhalid for more helpful content in practical.

— Hamza 💙