Switching from gpt-3.5-turbo-0613 to gpt-4o-mini

Community
1

Hello
I was recently sent an email that I should migrate away from my current gpt-3.5-turbo-0613 to the gpt-4o-mini

I am currently using the gpt-3.5-turbo-0613 for the Herika chat/GPT mod for the PC version of the game Skyrim, if you are familiar with the Herika mod. The current set up works great with quick and accurate response times, however when I migrated over to the gpt-4o-mini , the chat/GPT is either extremely slow or doesn’t respond at all. Should I be migrating to a completely different model for better performance?

I’m quite new to the chat/GPT scene. My current Herika server settings for the gpt-3.5-turbo-0613 are…
Connector openai max tokens 100
Connector openai presence penalty 1
Connector openai top_p 1
Connector openai Max Tokens Memory 512
Connector openai temperature 1
Connector openai frequency penalty 0

Should I perhaps be changing any of the above values when I migrate over to the gpt-4o-mini?

Your help is greatly appreciated! I have been having a great experience and no problems whatsoever so far with the current gpt-3.5-turbo-0613.

1 Like
2

Welcome to the forum.

Might just be network usage making it slower.

You could also try gpt-3.5-turbo but 4o-mini is usually faster.

Let us know. Would love to hear more about your experiences using an LLM for NPC dialogue…

1 Like
3

I’m already using gpt-3.5-turbo without any problems, but was told gpt-3.5-turbo would be shutting down Sept 13th and I should migrate to gpt-4o-mini, which I’ve done, but not working nearly as well or not at all. Any help?

4

Ah, I thought you meant this sub-model. They’re getting rid of gpt-3.5-turbo too? I noticed it wasn’t in ChatGPT anymore about a week ago…

5

Hi!

Where did you get the information that GPT-3.5 Turbo is about to be shut down? I double-checked the models page and couldn’t find any indication of such a step.

As of July 2024, gpt-4o-mini should be used in place of gpt-3.5-turbo, as it is cheaper, more capable, multimodal, and just as fast. gpt-3.5-turbo is still available for use in the API.

https://platform.openai.com/docs/models/gpt-3-5-turbo

Hope this helps.

1 Like
6

Yeah, it looks like only the -0614 sub-version is being deprecated…

2 Likes
7

if you are using gpt-3.5-turbo-0613, chances are the openai module you are using is probably old. if so, you can first try to update if you have not done yet.

8

Thanks guys! Yes, I believe it is in fact just the submodel 0613 that is going to be depreciated, so I might just try the regular gpt-3.5-turbo model. Hopefully that will work for me better than the gpt-4o-mini is working so far.

2 Likes