Anyone else can't update to 0125 (from 1106) without crashing their code's logic?

I’ve been using the 3.5 API for more than a year now. I’m building some seriously complicated logics on top of it, using it to both trigger various actions & use the text itself.
I used to update every time OpenAI released an updated version. I literally just changed gpt-3.5-X to gpt-3.5-X+1 or gpt-3.5 to gpt-3.5-turbo or even to -4 and it worked charm. My code’s logic remained untouched, and just produced better results.
It no longer does. I don’t know why, but when I changed gpt-3.5-turbo-1106 to gpt-3.5-turbo-0125, literally the day it was released, the code crashed. I guessed it’s just an issue with the version’s freshness, so I waited for it to stabilized. It still hasn’t.
For sure I can look into it, go through my logic & find where it fails. But updating is nice to have for me - surely not a must - and I have much more pressing things to work on other than looking back into my foundations for a slight upgrade.
I was just wondering: am I the only one facing this issue? Like, I guess many of us here build on top of this API & don’t restrcuture their logic when a new model is released. Furthermore, I would guess that the first thing OpenAI think of with regards to releases is backward compatibility. It’s the most fundamental part of releases. So is it just me having a very niche bug?

1 Like

I am having the same thing unfortunatly and worst is Assistant getting dummer in every update to be honest.


What are the specific errors or issues you are getting?

1 Like