Most of what I write to you is directly about fundraising.

This isn't, exactly. But I still think it matters for your fundraise, because the first part of fundraising is having a great business, and sometimes the highest leverage fundraising advice is actually product advice.

So this is a small detour.

If you have been meaning to play with Anthropic's new model, Fable, I think you should do it hard over the next few days.

Why? Because right now it is temporarily available inside Anthropic's subscription plans at a level that is, in my opinion, wildly underpriced relative to what it would cost through the API.

On the API, Fable is expensive. Like very very expensive…

Anthropic lists it at $10 per million input tokens and $50 per million output tokens. It also has a 1 million token context window, which means you can point it at huge codebases, giant product specs, ugly piles of data, or sprawling strategy docs without immediately running into the usual context ceiling.

Those numbers may sound abstract, so here's the simple version: if you use Fable seriously, it gets expensive fast.

A run with 200k input tokens and 50k output tokens is roughly $4.50.

A run with 500k input tokens and 100k output tokens is roughly $10.

Do that over and over while exploring a codebase, reviewing UX, iterating on strategy, or pushing on product ideas, and you can burn through a lot more than $200 pretty quickly.

That matters because Anthropic's current Max plan is $200/month, and for this promotional period they allowed subscribers to use Fable for up to 50% of their weekly usage limit. The original cutoff was July 7, 2026. Then, today, they extended it to July 12, 2026.

So if you have not started yet, you still have a few days.

At first I wasn't even sure I agreed with the internet on this one. Twitter was on fire about how good Fable was, and I was kinda like... ok, sure, but is it actually that much better?

I’ve always doubted that people could actually tell the incremental difference between each successive model that got released.  It all just felt so anecdotal “omg Opus is sooo good!”  but examples people gave were soft evidence at best.

Even with Fable I struggled to be convinced until I jumped into use. And now, I think I can confidently say the answer is yes…Fable is much better.

The cleanest example came from pointing it at the codebase for a little fun side project. 

Over the last 4 months, I’ve been picking at a fun little game called Timeline Frenzy. 

In theory, the game is pretty simple but in practice (and implementation) there are some challenging media streaming issues that both 5.5 and Opus struggled with…and the gameplay was intricate enough that UX design was complicated.  

Original Game Design

2 days ago, I pointed Fable at the Timeline Frenzy codebase with very little direction.  I basically just said “Review this, understand what's going on, and tell me where the real opportunities are to improve it.”

The response was impressive.  It came back saying, “I see that you are having problems activating streaming predictably. Your implementation is wrong. Here's the solution. Oh, and I've already done it.” 

That was impressive enough...

But then it helped reshape the design too.

In about two days of messing with it, the product went from something I was struggling to make truly valuable into something that feels much more like a real game. If you want to see the kind of product leap I'm talking about, go play with Timeline Frenzy at timelinefrenzy.com. It’s a real real robust game…

Fable materially helped me make that better. 

So my recommendation is pretty simple:

And if you're wondering what to use it for, here are the three best buckets I've found so far:

I'm not saying it will magically fix your business.

I am saying that if there is a temporarily underpriced way to aim a very powerful model at your company, and your company is what you are raising money for, this is probably worth your time.

The fundraising angle here is simple: better products, better strategy, and better execution make fundraising easier. If a few days of heavy AI usage can move those things forward, this is not a random detour from fundraising work. It actually is fundraising work.  And then this becomes not a random essay …but actually a fundraising essay :P 

So if you've been on the fence, get off it.

You have until July 12, 2026. tokenmaxxxxxxxxxxxx

Be chased, 
Jason

What's included: