How are file upload and inquiry operations priced in the API?

Below is an explanation based on the relevant pricing details for file search:

File Storage Cost

When you upload a file for operations (such as beiing used as a vector store for file search), the billing is based on the size of the underlying vector store rather than reading the file each time the inquiry is made.

• The file is first broken into chunks and embeddings are created.
• The first 1 GB of vector storage per day is free.
• Any usage above 1 GB is billed at $0.10 per GB per day.

For a 100 MB file, you are well within the free 1 GB limit. In other words, storing that file as a vector store will incur no daily storage costs.

Inquiry (Search) Costs

• The inquiry itself does not incur an additional token-based cost for “scanning” the file.
• The query operation works against the precomputed vector store. There is no extra charge (token-based or otherwise) for reading or scanning the stored data during an inquiry.

Conclusion

For the fictive scenario:

• 100 MB file (stored as a vector store): Falls within the free 1 GB daily allotment.
• One inquiry (search) into the file: Does not incur additional token or processing charges (apart from potential standard API usage where applicable).

So, as long as your overall vector store storage remains below 1 GB per day, a single inquiry into that 100 MB file will effectively cost nothing additional—which means there is no extra charge for scanning/reading the document during that inquiry.