RAG Cost Calculator
Separate the cheap one-time indexing bill from the recurring query bill.
Estimate your workload
Change the assumptions on the left; read the modeled result on the right.
How to read this estimate
What this calculator is for
RAG systems have two different economic shapes: indexing is mostly tied to corpus size, while generation scales with every user query. Keeping them separate shows what will dominate as traffic grows.
How to read the result
Ongoing monthly RAG cost combines generation and the vector layer, while one-time indexing is shown separately and not included in the monthly total. Cost per query divides ongoing spend by monthly queries. Changes to retrieved context length affect repeated generation cost.
The math
Ongoing monthly RAG cost = LLM generation for all queries + monthly vector/retrieval charges. One-time indexing is shown separately.
Pricing sources
Rates use the published TokenCOGS pricing registry and its September 16, 2026 fallback snapshot. Live values may differ; see methodology and source limitations before procurement.
Planning limitations
Actual invoices can differ because providers may apply long-context pricing, caching rules, batch rates, tool-call fees, regional processing, negotiated contracts, taxes, or future pricing changes. Treat the output as a planning estimate.
Use it to create scenarios and catch order-of-magnitude mistakes. Before committing spend, verify the selected model's live provider pricing and compare the estimate with actual usage telemetry.
Is the initial document-indexing bill included in monthly cost?
No. The one-time embedding estimate is separated because it depends on corpus size and reindexing frequency.
Is embedding usually the biggest RAG cost?
Often no. For active products, repeated generation over retrieved context can outweigh the one-time embedding bill, but your architecture and provider choices matter.