Is it possible to use only the gpt3.5 turbo model?

[bixinhaner]

I noticed that this project uses both models, Davinci-003 and GPT3.5 Turbo. What are their respective roles? As GPT3.5 Turbo is a general-purpose model, can we use only it to complete the working? Because the price of Davinci-003 is 10 times higher than GPT3.5 Turbo. 😄

[hophanms]

Davinci-003 gives better answer in Q&A

[nubgamerz]

@hophanms It is possible to use the davinci model instead of gpt-35-turbo? We've applied for gpt-35/gpt-4 but still on the waitlist. If we can utiliise davinci instead, it would be great.

[github-actions]

This issue is stale because it has been open 60 days with no activity. Remove stale label or comment or this issue will be closed.

[pamelafox]

Thanks for raising this back in April 2023.

Since then the sample has been updated to rely on a single Chat Completions ("chat") style model, configured via the environment variables AZURE_OPENAI_CHATGPT_MODEL and AZURE_OPENAI_CHATGPT_DEPLOYMENT. All references to the older dual‑model pattern (text-davinci-003 + gpt-35-turbo) have been removed, and there are no remaining "davinci" mentions in the current codebase. The original concern about needing both models (and the higher cost of text-davinci-003) is therefore no longer applicable.

Closing this as obsolete. If you run into a current scenario where using only a modern chat model (e.g. gpt-4.x / gpt-35-turbo or successors) doesn’t work with the latest version of the repo, please open a new issue with details of your configuration.

Thanks again for the original feedback!