RAG Query Cost Calculator
Break one RAG query into retrieval, rerank and generation cost.
Retrieval cost / query ($)
Rerank cost / query ($)
LLM cost / query ($)
Per-query = retrieval + rerank + LLM.
Live estimate
- Cost per query
- $0.00
- Cost per 1K queries
- $2.40
- Monthly at 60K queries
- $144.00
Per-query breakdown
LLM generation is 60–80% of per-query cost; retrieval is minor.
| Component | Typical | Share |
|---|---|---|
| Retrieval | $0.0003 | 13% |
| Rerank | $0.0002 | 8% |
| LLM (1.9K tok) | $0.0019 | 79% |
Frequently asked questions
- How is rag query cost calculated?
- Per-query = retrieval + rerank + LLM. LLM generation is 60–80% of per-query cost; retrieval is minor.
- What inputs do I need for the rag query cost calculator?
- Enter retrieval cost / query ($), rerank cost / query ($), llm cost / query ($). Defaults reflect typical 2026 usage; replace them with your own vendor quote or last month of logs for an exact figure.
- What is a typical rag query cost benchmark?
- A standard RAG query costs ~$0.0024 — $144/month at 60K queries.
- Do I need reranking?
- Rerank adds ~8% cost but often lifts accuracy more than retrieving 2× passages.
- How can I lower my rag query cost?
- Cut volume waste first (shorter durations, fewer retries, tighter prompts), then route to cheaper tiers for drafts, enable caching/batch where available, and cap per-user or per-job spend with alerts. Re-run the calculator after each change to size the saving.
- Is this calculator free?
- Free forever, no sign-up, runs 100% in your browser. Copy results or export CSV for your budget doc.