What is cost per query?
Cost per query is the total cost of a single end-to-end interaction with your AI application — including all model calls, tool use, retrieval, and retries. It is always higher than a single API call because real applications make multiple calls per query.
A simple chatbot: 1 call. A RAG application: 1 embedding call + 1 vector search + 1 generation call = 3 calls. An agent: 3–15 calls depending on task complexity. If each GPT-4o call costs $0.003 and your agent averages 8 calls per query, your cost per query is $0.024.
Why it matters
Per-million-token pricing tells you the unit cost. Cost per query tells you the operational cost. The gap between them is your application architecture. sourc.dev tracks per-million prices because they are the standard comparison unit — but cost per query is what determines whether your product is viable at scale.