# How much does Assistants API Retrieval Pricing costs?

**URL:** <https://community.openai.com/t/how-much-does-assistants-api-retrieval-pricing-costs/510003>\
**Category:** API\
**Created:** [November 17, 2023, 2:55pm UTC](https://community.openai.com/t/how-much-does-assistants-api-retrieval-pricing-costs/510003 "2023-11-17T14:55:26Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![paul86](https://avatars.discourse-cdn.com/v4/letter/p/d9b06d/32.png) [@paul86](https://community.openai.com/u/paul86)\
**Post date:** [November 17, 2023, 2:55pm UTC](https://community.openai.com/t/how-much-does-assistants-api-retrieval-pricing-costs/510003/1 "2023-11-17T14:55:26Z")

</div>

Hello, it is unclear to me what happens if I upload a file with e.g. 100,000 words for specialist knowledge. What exactly do I have to pay for? Also for reading the tokens that correspond to the 100,000 words?

---

<div class="post-metadata">

**Author:** ![Foxalabs](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/foxalabs/32/738355_2.png) [@Foxalabs](https://community.openai.com/u/Foxalabs)\
**Post date:** [November 17, 2023, 3:20pm UTC](https://community.openai.com/t/how-much-does-assistants-api-retrieval-pricing-costs/510003/2 "2023-11-17T15:20:06Z")

</div>

You will have to pay $0.20 per day per gigabyte of stored information or part thereof, plus the API token count used by the retrieval system as context.

---

<div class="post-metadata">

**Author:** ![\_j](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/_j/32/766292_2.png) [@\_j](https://community.openai.com/u/_j)\
**Post date:** [November 17, 2023, 3:31pm UTC](https://community.openai.com/t/how-much-does-assistants-api-retrieval-pricing-costs/510003/3 "2023-11-17T15:31:03Z")

</div>

Let’s say that you don’t attach any functions, code interpreter, and ask in a new thread. Minimize the autonomy that the agent backend has to iterate on you. The API documentation basically gives a promise that the model context will be filled with retrieval (and no max\_token to set).

How about 100k+ tokens of context loaded up for gpt-4-turbo = $1.00? Your 100000 words = around 150k tokens to be embedded and chunked back into the model context.

You’ll only get a delayed daily report of the tokens used in total for the day - by design, I’m sure, with the obfuscated billing usage coming out in concert with assistants.

---

<div class="post-metadata">

**Author:** ![trenton.dambrowitz](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/trenton.dambrowitz/32/146436_2.png) [@trenton.dambrowitz](https://community.openai.com/u/trenton.dambrowitz)\
**Post date:** [November 17, 2023, 4:04pm UTC](https://community.openai.com/t/how-much-does-assistants-api-retrieval-pricing-costs/510003/4 "2023-11-17T16:04:12Z")

</div>

I’m rather upset at the fact that they removed the daily breakdown.

It would be extremely helpful to see what I used and when I used it. Instead we get an aesthetically pleasing but functionally useless line graph.

---

<div class="post-metadata">

**Author:** ![grandell1234](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/grandell1234/32/286916_2.png) [@grandell1234](https://community.openai.com/u/grandell1234)\
**Post date:** [July 7, 2024, 5:32pm UTC](https://community.openai.com/t/how-much-does-assistants-api-retrieval-pricing-costs/510003/5 "2024-07-07T17:32:31Z")

</div>