I have fine-tuned the base models on bedrock in order to use the fine-tuned model we have to deploy the model and there is a Provisional Throughput cost, but never on OpenAI, so want to know is there any Provisional Cost or Deployment Cost of fine-tuned model.
What is the provisional, deployment and storage cost of fine-tuned model? Is there any commitment period for Fine-Tuned model?
Hi and welcome the community!
OpenAI charges only for training and using the fine-tuned model.
All other costs you are referring to are AWS related.
Edit: Found this clear statement on the pricing page
Once you fine-tune a model, you’ll be billed only for the tokens you use in requests to that model.
Just one more query the usage cost, is the Input and Output cost mentioned per 1000 tokens on the Pricing Page of fine tuning model right?
Yes, I just edited the reply above to include that information.
This is likely the part of the table that matters to you the most:
You can get the token counts via tiktokken: