Okay this one’s wild.
On June 9th Anthropic launched Claude Fable 5. Their most capable model ever. Three days later the US government ordered it offline under export controls. Nobody could touch it for three weeks.
As of July 1st it's back.
And here's what matters to you right now. If you're on a Claude Pro, Max, or Team plan, Fable 5 is sitting in your account included in your subscription until July 7th.
Up to 50% of your weekly usage limit at no extra cost.
After Monday it moves to paid credits. On the API it already costs $10 per million input tokens and $50 per million output. That's double their previous top model and the most expensive thing Anthropic sells.
So you have six days of free access to a model that costs a lot next week.
Most people will spend it rewriting emails and summarizing docs. That's like renting an excavator to plant flowers.
Here's the one thing you need to understand about Fable 5 before you use it.
Its edge isn't speed. It isn't writing quality. It's this: the longer and more complex the task, the bigger its lead over everything else.
Stripe used it to compress a codebase migration across 50 million lines of code into a single day. A project their team had estimated at two months.
You don't have 50 million lines of code. But you have a version of that problem. Every business does. It's the project you keep breaking into 20 smaller pieces because no tool could hold the whole thing at once.
That's exactly what this week is for.
Don’t give it tasks, give it problems.
A task is "write me 5 hooks." A problem is "here's my last 90 days of content, my analytics, and my offer tell me why conversion is flat and rebuild the whole system to fix it." Completely different animal.
One caveat before you dive in. The redeployed version has stricter security guardrails. If your work touches code debugging or anything security adjacent, expect the occasional refusal.
Here are prompts you can steal right now.
Swap in your details, dump as much real context as you can: files, exports, transcripts. Context is the fuel.
If you run an agency
Here are briefs, content calendars, and performance data for 10 of my clients. Analyze them as one portfolio. Find the patterns, which hooks, offers, and formats consistently underperform across clients and why. Give me one systemic fix per client, ranked by expected revenue impact.
Here's my service list, current pricing, delivery costs in hours, and my last 12 months of proposals with win/loss outcomes. Rebuild my pricing structure. Show me where I'm underpriced, which services to kill, and what a productized tier would look like. Include the math.
If you sell products
Here's my order export from the last 18 months and my current inventory. Build my Q4 launch plan demand forecast by SKU, inventory order timing, discount structure that protects margin, and a week-by-week email and SMS calendar from October 1 to December 1.
Here are 200 customer reviews and support tickets. Extract the exact language customers use to describe the problem my product solves. Rewrite my product page using their words not mine. Flag the top 3 objections and give me an FAQ that kills each one.
If you're a creator or coach
Here are my last 60 posts with performance numbers. Don't tell me what performed. Tell me why, find the pattern in hooks, topics, and formats, then design a 30-day content system built only on what the data says works. Include 30 hooks written in my voice.
I sell [your offer] at [price]. Here's my audience size, engagement data, and what my followers DM me about. Design my full offer ladder from free lead magnet to premium, with pricing logic, and tell me which piece to build first and why.
If you run a local service business
Here's my booking data for 6 months including cancellations and no-shows. Find the pattern which days, services, and customer types flake most. Design a reminder and deposit system to cut no-shows, with the exact SMS scripts.
Works for any business
Here's everything about my business what I sell, my numbers, my team, my constraints. Act as my strategy team. Identify my single biggest bottleneck, prove it with my own data, and build the 90-day plan to break it. Challenge my assumptions. I want the plan you'd give me if you were paid to disagree with me.
A few things before you go use this.
Pick two or three problems maximum. The 50% usage cap is real and Fable burns through it faster than other models.
Spend it like a budget not a buffet.
Front-load your context every time.
Exports
Transcripts
Briefs
Numbers.
The model is only as good as what you hand it.
Route the small stuff to standard models. Captions, emails, quick edits, save Fable's allowance for the heavy lifts.
And save everything it builds you. Whatever Fable produces this week you keep forever. The systems, plans, and frameworks don't expire on July 7th, only the free access does.
Nav
