While exploring the OpenAiApi class, I identified some areas for improvement within the ChatModel and ChatCompletionFinishReason enums and proceeded with the following adjustments:

  • I updated the version in ChatModel from gpt-3.5-turbo-1106 to gpt-3.5-turbo-0125.
  • I corrected a typo in ChatCompletionFinishReason, changing TOOL_CAL to TOOL_CALL.

Comment From: markpollack

I think that since the gpt-3.5-turbo will always point to the "unpinned" gpt-3.5-turbo model. However, users may still want to explicitly specify by name the model they want to use, even if at some point in time it matches the model that is the alias for gpt-3.5-turbo model. I'll update the PR to add the new version instead of replace the older one.

Comment From: youngmoneee

I think that since the gpt-3.5-turbo will always point to the "unpinned" gpt-3.5-turbo model. However, users may still want to explicitly specify by name the model they want to use, even if at some point in time it matches the model that is the alias for gpt-3.5-turbo model. I'll update the PR to add the new version instead of replace the older one.

I've made modifications to keep the existing version without removing it

Comment From: markpollack

merged in 7396597678184a1a555cf1724fb65bb354e35152 Thanks!