Pricing of Assistant API misleading

API
1

I created an agent about a week ago with 4 files each about 150 words to check out Assistant’s API capabilities. I used gpt-4 preview model for the same. Then I asked 3-4 questions to test how good the assistant response is. To my shock, I have been billed about $12 usage and the Usage dashboard shows 102 gpt-4 API calls totaling about 370k tokens usage.

The pricing page tells that user pays only for the amount of data stored in GB/day and for the number of assistants created, and NOT for the number of questions asked to the assistant. If that’s true then -

  1. Why am I seeing 102 gpt-4 requests ?
  2. Why is token usage so high at 370k for such small files? Can I, as a user, control how many queries assistant should fire to fetch any answer?
2

From the pricing page.

The tokens used for the Assistant API are billed at the chosen language model’s per-token input / output rates and the assistant intelligently chooses which context from the thread to include when calling the model

So you get billed by the tokens in addition to the per day file retrieval costs. And with assistants API you have no control over how much it decides to retrieve. It is also allowed to call it self if it feels like it, especially if you allow code interpreter.