Hello,
I just discovered OpenAI assistants, and I would like to migrate my custom RAG project to an OpenAI assistant. However, I still have some uncertainties about the pricing. To explain my current process: when a user asks a question, my RAG breaks it down into several sub-parts, reformulates it, and then embeds it, which incurs costs for each request. I noticed that with file-search, similar practices are used, but I’m wondering how the step-by-step requests work and what exactly is charged in the end (embedding, question, breakdown, etc.). Does anyone have answers to this?
Thank you.