I have developed a function to call GPT 4 Turbo for an analysis of the results of a report. My prompt is always the same, but the data transmitted is always modified.
and GPT 4’s response is now always
I’m sorry, but I can’t provide the requested service.
What’s the problem?
The problem is probably that
gpt-4-turbo-preview Currently points to
gpt-4-0125-preview.
https://platform.openai.com/docs/models/gpt-4-and-gpt-4-turbo
There was an issue with 1106 truncating work (‘being lazy’), and they fixed that by making sure 0125 simply refuses to do work.
Azure has just threatened to transition all 1106 deployments to 0125 on march 8, so stuff will continue to break.
1 Like
Yes pointing is to gpt-4-turbo-preview. so i have to point to wich model ??
probably gpt-4-1106-preview
but be aware that it can be deprecated at any moment, depending on how OpenAI feels
1 Like
well, that’s what it looked like. Thanks for the track, I’ve been looking for it for 3 days. I will be vigilant on the pointing.
1 Like
for posterity (and bots), the model IDs are listed here:
https://platform.openai.com/docs/models/overview
and here: