The original Fable 5 was one of the best models I ever dropped into a real workflow. Then it got shut down, the second version showed up to replace it, and that one just is not the same. That is basically the whole post. But there is a real lesson buried in here about how I actually decide what earns a spot in my stack, so let me walk through it.
So here is where I land. Until Fable 5 gets folded into the full Claude Max plan, I am not setting it up. Not because it is bad, the first one was legitimately great. Because the juice is not worth the squeeze when it lives off on its own as a separate thing I have to wire up, babysit, and pay for on the side.
The first Fable 5 was the real one
That first version surprised me, and I do not surprise easy anymore (you use enough of these things and the shine wears off fast). It was fast, it held a thread across a long agent run without wandering off, and it just felt like it got what I was trying to do. I had it doing real work inside a couple of my agent loops within a day.
And then it was gone. Ok, that happens, models get deprecated, I get it. But you do not forget what the good version felt like.
The second one is not it
The replacement is fine. And fine is exactly the problem, right. When something is genuinely great you will jump through hoops to keep it. When it is just fine, every little bit of friction suddenly counts against it, and there is a lot of friction in standing up a whole separate model outside the tools I already run.
I tried it (I really did, I am not just talking) and I kept catching myself thinking I would rather just be back in Claude. That is not a review, that is a gut check, and my gut is usually right about this stuff.
Why Fable 5 stays on the shelf for now
Here is the math nobody tells you about new AI tools. Every one you add has a tax. Setup, keys, another bill, another dashboard, another thing that breaks at 2am and pages you while your agents sit there stuck. A tool has to clear that tax before it earns its keep, and a merely-fine model does not clear it.
Claude Max is already where my agents live. If Fable 5 showed up inside that plan, no separate setup, no separate bill, I would have it running by lunch and I would tell you how it went. But as a standalone thing I have to go out of my way for, to get a version that is a step down from the one I loved. Yeah, I cannot be bothered. Dan Harris’s whole thing in 10% Happier is being the skeptic in the room who makes you prove it, and that is roughly my posture on any tool that wants a slot in my stack. Prove it is worth the tax or stay on the shelf.
So thats my hot take. Fold Fable 5 into Claude Max and I am in. Until then it is a hard pass, and I am ok being wrong about it later.
Be good to people.