Is there a per-call cost for the direct Vector Store Search API?

[ndhoanit1112]

I’m trying to understand the pricing for the direct Vector Store Search API using the Python SDK:

client.vector_stores.search(
    vector_store_id="vs_abc123",
    query="your query"
)

According to the official documentation (Assistants API > Tools > File Search > Managing costs with expiration policies), it states:

“You first GB is free and beyond that, usage is billed at $0.10/GB/day of vector storage. There are no other costs associated with vector store operations.”

However, I’ve seen some users in this forum report being billed for “file search tool calls” even when using the direct Vector Store Search API (not through the Responses API).

My question:

Is there a per-call/per-query cost when using client.vector_stores.search()?

If yes, what is the rate? Is it the same as the Responses API file_search tool ($2.50 per 1,000 calls)?

Thanks in advance for any clarification!

[_j]

Costs, here:

Vector Store **Search** API endpoint, the missing documentation

Quickstart Guide: Using OpenAI’s Vector Store Search Endpoint Forward I had to gather all the resources into one to power this AI assisted compilation… I don’t care enough right now to actually go after making a robust database to search to offer more practicality and example… Introduction OpenAI’s Vector Store Search Endpoint enables developers to query and retrieve highly relevant document chunks from a custom vector store hosted within OpenAI’s API infrastructure. This endpoint is powerful…

Same per-use rate as Responses. When you call the endpoint yourself, you incur the single-use cost. If you use an AI with the file_search tool, it might decide to make multiple “uses” for you..

The big cost of an embeddings-based vector store is paying the up-front costs of the ingesting of files. Then after that, the single embeddings query calls are very cheap. OpenAI switches this up with low barrier, where they recover their costs of providing you no-cost vector store creation with daily storage fees, per call fee, and the large context placement into an AI model if using as tool.

Assistants didn’t have a usage fee for file search, but of course you didn’t have direct access to the results nor could you write your own query. It’s being shut off next year.

[ndhoanit1112]

Thanks for your reply!

I haven’t been able to confirm this on my end yet. When I tested client.vector_stores.search() myself, I couldn’t find where these calls are reflected in my API usage dashboard.

I checked:

• Usage > File Searches — shows 1 request, but this was from a Responses API file_search call, not from the direct client.vector_stores.search() call
• Vector Stores — shows 0 bytes (storage only, no call count)

Could you point me to which usage metric reflects the direct Vector Store Search API calls (client.vector_stores.search())?

I want to make sure I’m tracking the costs correctly.

Thanks!

[_j]

Here’s your link (including a two-day range of “now”):

https://platform.openai.com/usage/file-searches?startDate=2025-12-18&endDate=2025-12-19

Here’s the search API calls I just made showing up with their count:

Return back to the main “usage”, group by “line item” - I get to pay for them. $0.008

[ndhoanit1112]

You’re right! I can confirm now.

The issue was that I was filtering by API key in the usage dashboard. When I cleared the filter, the calls showed up correctly. Strangely, filtering by API key only shows Responses API file_search calls, not direct client.vector_stores.search() calls.

Thanks for your help in clarifying this!