Applications that use the stable model names for the GPT-3 base models (ada, babbage, curie, davinci) will automatically update to the new models listed above on January 4, 2024. New models will also be accessible in the coming weeks for early testing by specifying the following model names in API calls: ada-002, babbage-002, curie-002, davinci-002.
Developers using older completion models (such as text-davinci-003) will need to manually update their integration before January 4, 2024 by specifying gpt-3.5-turbo-instruct in the “model” parameter of your API requests. gpt-3.5-turbo-instruct is an InstructGPT-style model, trained similarly to text-davinci-003. This new model is a drop-in replacement in the Completion API and will be available in the coming weeks for early testing.
Developers who wish to continue using their adjusted models beyond January 4, 2024 will need to adjust the replacements on top of the new GPT-3 base models (ada-002, babbage-002, curie-002, davinci-002), or newer models (gpt-3.5-turbo, gpt-4). Once this feature becomes available later this year, we will give priority access to fine tuning the GPT-3.5 Turbo and GPT-4 to users who have previously tuned older models. We recognize that migrating from models that are fitted to your own data is challenging. We will support users who previously tuned the models to make this transition as smooth as possible.
In the coming weeks, we’ll be reaching out to developers who have recently used these older models with more information once the full new models are ready for early testing.
