I set this up on a new machine today.
I realised that in 2 games I had suddenly spent 7 dollars instantly. Pretty steep considering it was supposed to be using 3.5-turbo-16k
Checked over all the setting. Then parsed the entire source code for ANY reference to default models. anything. Changed every single reference. No matter what I set it is using the GPT4 API endpoint.
I have tried; the usual inline declarations; using the built in preferred template; just jamming it in as the hardcoded value to the method that checks it. So.. unless anyone has any other ideas. I'm all out.
7 bucks is no big shakes but I run systems in parallel and no doubt others batch too. Wouldn't want anyone to get an AMAZON AWS level surprise and not have reported. so. Enjoy. Reported.
EDIT: It's a linux box but I did install it under windows on there as well and test it (that test is probably the top 50c of the green bar ) It was identical so "not OS". and "yes, I also tried that."
I set this up on a new machine today.
I realised that in 2 games I had suddenly spent 7 dollars instantly. Pretty steep considering it was supposed to be using 3.5-turbo-16k
Checked over all the setting. Then parsed the entire source code for ANY reference to default models. anything. Changed every single reference. No matter what I set it is using the GPT4 API endpoint.
I have tried; the usual inline declarations; using the built in preferred template; just jamming it in as the hardcoded value to the method that checks it. So.. unless anyone has any other ideas. I'm all out.
7 bucks is no big shakes but I run systems in parallel and no doubt others batch too. Wouldn't want anyone to get an AMAZON AWS level surprise and not have reported. so. Enjoy. Reported.
EDIT: It's a linux box but I did install it under windows on there as well and test it (that test is probably the top 50c of the green bar ) It was identical so "not OS". and "yes, I also tried that."