Fable 5.1 is incredible.

@theo You should take your time though, last time you said the similar things about opus 5 and praised it a lot then bashed it even more after that lol So tell us when you are sure about fable 5.1

This video is me explaining how I did this correctly and learned my lesson about Opus 5

oh thats good then I will watch it now for sure

It’s only good if you have the money to burn on 5 fable accounts.

You said the same thing about Opus 5, then spent weeks tearing it down. Now Fable 5.1 is "incredible" before the paint is dry. The consistent part isn’t the model. It’s the hot take. You are consistently .. inconsistent !! 🤷‍♂️

Theo 12 hours later :

Then why does T3 code still feel like slop?

i just said “Hi” to Fable 5.1 and it was already about $0.30 😅

Sure… 👌

If you have the money to use, Fable 5.1 is the best model right now. In my experience though, it burns through the usage limits way too fast...

incredible in token usage

I am going to wait for Astra still. Codex is the goat and price effective. Fable sure is good but at what cost and how much better is it really 5%, 10%?

Simping for misAnthropic again. It’s garbage actually. Hallucinating more than opus 5, watermarking your every move and wasting your money like never before.

here I am quitting my agents because they spawn children on simple questions while this guys blows generational wealth into a bunch of numbers that seem to get bigger and better

Incredible on the tape. The other number is the window. Same model can look like this and still empty Max before the task ends.

the thing worth checking in 5.1 is whether it still spins up parallel agents when the task doesn't call for it, that was the real complaint on 5.0, not raw output quality.

low iq bs. soon nobody is going to cate about anthropic

Yes, but it still does dumb shit and sometimes does not follow the prompt. I’m getting Opus 5 vibes from it. I’m most impressed with how it tries to debug things, but only if you order it to. It never did it on its own for me. Here is a screenshot of the animation debug it did.

Model's good but company sucks. Anthropic hardly cares about its non enterprise customers

Fable 5.1 feels like a step back to me. Its outputs are more verbose and less clear. UI/UX feels slightly harder to steer. Ideation feels harder. And of course, usage is terrible. I could be wrong, but this has been my experience as someone using Fable 5 for the last three months. Sucks that we can't use Fable 5 in Claude Code anymore.

Mythos?

Where do we find this 'unslop' skill?

@theo Seems like you get to test multiple models (Astra 6, Fable 5.1) before release. Are the number of PRs you've been merging a mixed bag here, or only Fable 5.1?

is it incredible like opus 5 was? imma wait 2 more weks to listen to your opinion :))

17 seconds ago is crazy

crazy acceleration loop Next layer is taste and worth doing it at all

gpt 6

Fable's turnaround is making my release cadence look like a government procurement timeline.

Wow, that's awesome! I'm glad you're enjoying Fable 5.1 so much. It really is something special.

Fable 5.1 is genuinely impressive. That said, the pricing, token consumption, and subscription tiers could use some rethinking.

Time to pin a new video 👀

Fable 5.1 is cracked

Next video fable 5.1 is terrible next video gemini 3.8 flash is the most flagship model next video glm is what I use all day

It’s burns tokens so fast it’s essentially unusable

Yes incredible indeed. Astra cannot come soon enough.

Incredible is fine. I still want the cache-write line on the bill.