Reference table
Embeddings DB cost by stored vectors and query volume
Short answer: for a RAG index the monthly bill is driven by how many vectors you store and how many searches you run. Across this 42-cell sweep of 768-dim vectors, the cheapest hosted option ranges from $16.00 to $3,637.14/mo, and Turbopuffer is the cheapest hosted option in every cell. At 10M vectors and 250k queries/day it is $247.97/mo on Turbopuffer. Running LanceDB on your own object storage costs less, $2.86 to $29.66/mo for storage and requests, plus the compute you run it on. The full grid is below.
Cheapest hosted monthly cost by stored vectors and queries per day
Rows: stored vector count. Columns: queries per day. Each cell is the cheapest monthly cost among the hosted options (Pinecone, Postgres+pgvector on Supabase, Turbopuffer), with the winning vendor noted. Green < $25, unshaded $25–$100, amber $100–$500, red ≥ $500.
| Vectors \ q/day | 10k | 50k | 100k | 250k | 500k | 1M | 2M |
|---|---|---|---|---|---|---|---|
| 1M | $16.00Turbopuffer | $16.00Turbopuffer | $16.00Turbopuffer | $31.49Turbopuffer | $58.28Turbopuffer | $111.86Turbopuffer | $219.02Turbopuffer |
| 5M | $16.00Turbopuffer | $32.55Turbopuffer | $56.34Turbopuffer | $127.71Turbopuffer | $246.66Turbopuffer | $484.56Turbopuffer | $960.36Turbopuffer |
| 10M | $23.19Turbopuffer | $60.65Turbopuffer | $107.48Turbopuffer | $247.97Turbopuffer | $482.12Turbopuffer | $950.42Turbopuffer | $1,887.02Turbopuffer |
| 25M | $41.47Turbopuffer | $91.22Turbopuffer | $153.41Turbopuffer | $339.98Turbopuffer | $650.93Turbopuffer | $1,272.83Turbopuffer | $2,516.63Turbopuffer |
| 50M | $70.19Turbopuffer | $133.46Turbopuffer | $212.55Turbopuffer | $449.80Turbopuffer | $845.23Turbopuffer | $1,636.09Turbopuffer | $3,217.81Turbopuffer |
| 100M | $122.72Turbopuffer | $193.36Turbopuffer | $281.67Turbopuffer | $546.57Turbopuffer | $988.08Turbopuffer | $1,871.10Turbopuffer | $3,637.14Turbopuffer |
How the table is made
- Engine
- Embeddings DB Cost (
embeddings-db-cost) - Source
- Computed live from
/engines/embeddings-db-cost.js - Grid
- 6 vector counts × 7 query volumes = 42 cells
- Computed
- , recomputed on every build
- Held constant
- 768-dim vectors; 20,000 ingests/day; list pricing for Pinecone, Supabase (pgvector), Cloudflare R2 (LanceDB) and Turbopuffer, all 2026-10, with the Pinecone $50/mo and Turbopuffer $16/mo plan minimums modeled
Every value above and in the table is the deterministic return value of the shipped engine bundle, recomputed independently in continuous integration and diffed against this page on every build. The engine is pure (no clock, no randomness): the same inputs always produce the same output. No number on this page was hand-typed or estimated.
How to read this dataset
- Find your stored vector count on the left, your daily query volume across the top, and read the cheapest monthly cost and winning vendor where they meet.
- LanceDB on object storage is the lowest line in every cell, but its figure is storage and requests only; add the compute that runs your queries before comparing it with the hosted options.
- The hosted options scale on different meters. Pinecone bills each query by index size, from $50.00 in the smallest, quietest cell to $343,447.97 in the largest, busiest one. Postgres with pgvector on Supabase is priced by the compute size that holds the HNSW index in memory, from $225.00 at 1M vectors to $26,207.41 at 100M, almost regardless of queries. Turbopuffer's per-query charge falls steeply for large namespaces.
- The CSV carries every vendor's full monthly price per cell plus the index storage in GB, so you can compare like-for-like and add your own compute overhead.
Want to run your own inputs instead of this grid? Use the Embeddings DB Cost calculator. For the exact math, defaults, and what the model can't tell you, read the methodology page. For a head-to-head on two vendors, see Pinecone vs Qdrant pricing, and for the build-vs-buy call, managed vector DB vs. self-hosted pgvector.