I have a 10K tokens prompt,. Does gpt-4-turbo handle that ? Some times gpt-4-turbo generate really good aswers but some times it get confuse and I was reading that gpt-4-turbo onlye support 8K tokens is that correct? I asked to gpt-4o and its told me that I should use gpt-4-32k but that model does not listed in my dashboard. I am not sure if that model really exist
1 Like
You can look at the context window of models here:
https://platform.openai.com/docs/models#gpt-4-turbo-and-gpt-4
Also observe real model names instead of ones that the AI made up (or to which no new users are granted access, like 32k)
That context window is the combined area for placing input and forming output, so for gpt-4-turbo at 125k, you could subtract its maximum output of 4k and still have 121k tokens for “prompt”, which is messages and tools.
2 Likes