First initial thoughts of Fable 5.
It’s extremely intelligent and detailed , super impressed with the response quality and reasoning. Like this shit is insane.
The amount of depth it went into and the approach it had and stuck to was phenomenal, design wise was also significantly better.
Compared it all to GPT 5.5.
On Claude Code:
I used one monster audit prompt for a small project I was vibe coding ( I’ll put the prompt in the comments section, S/O to the X user who gave it )
But after that one prompt, the result blew me away, I used Fable 5 ultracode mode. The depth and what the audit report came back with was truly amazing.
But it also blew my entire 5 hour limit and used 15% of my weekly limit ( one shotted )
Final thoughts: it is a phenomenal model, but sheesh this thing is expensive. We’re going to have to figure out how to make these models cheaper for use, otherwise it’s not sustainable.
The price for value is so off.