Which models to use for text generation

API
1

Need 3 models for text generation, but unsure which to use.
Need one that is Good and cheap
one that is best for price (will be my default one)
and one that gives the best text (but may be more expensive to use)

I have access to these models:
gpt-4o-realtime-preview-2024-12-17
gpt-4-1106-preview
tts-1-hd
omni-moderation-2024-09-26
text-embedding-3-large
gpt-4o-mini-realtime-preview-2024-12-17
gpt-4o
gpt-4o-realtime-preview-2024-10-01
o1-mini
gpt-4o-audio-preview-2024-10-01
gpt-4-0125-preview
gpt-4o-audio-preview-2024-12-17
gpt-4-turbo-preview
gpt-4-turbo
gpt-3.5-turbo-0125
chatgpt-4o-latest
o1-preview
text-embedding-ada-002
gpt-4o-mini-realtime-preview
gpt-4o-2024-05-13
gpt-4o-mini
tts-1
davinci-002
dall-e-3
gpt-3.5-turbo
gpt-3.5-turbo-instruct-0914
tts-1-1106
gpt-4o-2024-08-06
gpt-3.5-turbo-instruct
gpt-3.5-turbo-1106
babbage-002
whisper-1
gpt-4o-mini-2024-07-18
gpt-3.5-turbo-16k
gpt-4o-mini-audio-preview-2024-12-17
gpt-4o-2024-11-20
gpt-4
gpt-4o-audio-preview
gpt-4-turbo-2024-04-09
text-embedding-3-small
tts-1-hd-1106
o1-mini-2024-09-12
o1-preview-2024-09-12
omni-moderation-latest
gpt-4o-realtime-preview
gpt-4-0613
dall-e-2
gpt-4o-mini-audio-preview

1 Like
2

Here you go…

Need one that is Good and cheap

gpt-3.5-turbo

one that is best for price (will be my default one)

gpt-4-turbo, more expensive but better

gives the best text (but may be more expensive to use)

gpt-4

1 Like
3

Thanks… wasn’t sure if some of the O’s or Mini’s was better choices. There were so many to choose from.

4

Hi @zarkas and welcome to the community!

I can also add the following:

  • Check the pricing on the models because some of the older models like gpt-4-turbo are more than double the cost of the latest gpt-4o
  • Keep in mind technical constraints like context size - older models like gpt-3.5 have significantly smaller context windows, meaning you would need to supply lower word count on the input, and would get lower word count on the output
  • There are some additional features such as prompt caching and structured outputs that are available on the newer models like gpt-4o, that will save you money, provide fast performance, and give an overall more robust development experience
  • gpt-4o in particular has been pinned as the most “creative” of all the models
1 Like