Embedding Cost Calculator

Estimate the cost to embed a corpus of text.

Quick answer

With the default inputs, $0.0200 — 1,000,000 tokens x 1 pass(es) with text-embedding-3-small. Enter your own numbers below to recompute instantly; the full step-by-step math is shown under the worked example.

Embedding model

Total tokens to embed(tokens)

Number of passes

Results update automatically as you type.

Result

$0.0200

1,000,000 tokens x 1 pass(es) with text-embedding-3-small

Tokens: 1,000,000
Rate: $0.02/M
Passes: 1
Total cost: $0.0200

Estimate the cost of generating embeddings for your text. Enter the total token count and how many times you will re-embed, and pick an embedding model to see the total cost.

How this is calculated

Cost = total tokens ÷ 1,000,000 × the embedding model's per-million price, multiplied by the number of passes over the corpus. Embeddings are input-only, so there's no output charge — the whole cost is the tokens you embed.

Is this a good result? What to do next

Embedding is usually the cheapest part of a RAG pipeline — cents to a few dollars for millions of tokens. The number matters mainly when you re-embed a large corpus repeatedly or move to a pricier, higher-dimension model.

Typical planning ranges

Small embedding models: cheapest per 1M tokens
Large embedding models: pricier, higher dimensions
One pass over 1M tokens: typically well under $1

Ranges are typical planning figures to sanity-check your result, not authoritative benchmarks. Your numbers will vary with use case, volume, and vendor.

How to improve this number

Use a smaller embedding model where retrieval quality holds.
Embed once and avoid unnecessary re-embedding.
Deduplicate and clean text before embedding.

Common mistakes

Re-embedding the whole corpus when only some docs changed.
Assuming a bigger embedding model is always better.

When to use a different approach

For recurring refresh cost, use the knowledge base refresh calculator. For storage of the resulting vectors, use the vector database storage calculator.