Fine-Tuning Cost Calculator
Estimate what a fine-tune costs to train and run, and when it pays for itself.
Embedding Cost Calculator
Estimate the cost and vector storage of embedding a document corpus for RAG.
Your corpus
Every month
Quick Tool Overview
The Embedding Cost Calculator estimates chunks, vectors, embedding tokens, one-off and monthly cost, and vector database storage for a retrieval corpus, across OpenAI, Cohere, Voyage, Google and open embedding models at float32, half, int8 or binary precision.
- What it does
- Estimates embedding cost and vector storage for RAG
- You give it
- Documents, length, chunking and monthly activity
- You get back
- Vector count, costs and storage to plan for
- Best for
- Planning semantic search and RAG systems
About the Embedding Cost Calculator
Building retrieval over your documents means splitting them into chunks, turning every chunk into an embedding vector, storing those vectors, and embedding every search query as it arrives. This calculator estimates each part before you commit to a model or a database.
Enter the number of documents and their average length, then the chunk size and overlap. You get the chunks per document, the total vectors, and the tokens embedded — including the overlap, which is embedded twice. Choose an embedding model from OpenAI, Cohere, Voyage, Google or an open model to load its price and dimensions, and add how much of the corpus changes each month and how many queries you serve.
The storage estimate is often the more important number. Raw vectors, an HNSW index and the chunk text are sized separately, and switching the stored precision from float32 to halfvec, int8 or binary shows how much memory your vector database server really needs.
Key features of the Embedding Cost Calculator
Chunking maths included
Chunks per document and the extra tokens embedded twice by overlap.
One-off and monthly cost
Initial embedding plus re-embedding churn and query embeddings.
Storage breakdown
Raw vectors, HNSW index and chunk text sized separately.
Precision options
float32, halfvec, int8 and binary, to see how much memory quantized vectors save.
Popular embedding models
Prices and dimensions load automatically and stay editable.
How to use the Embedding Cost Calculator
-
1
Choose an embedding model
Its price per million tokens and vector dimensions fill in automatically.
-
2
Describe the corpus
Number of documents, average words per document, chunk size and overlap.
-
3
Add monthly activity
The share of the corpus re-embedded each month, and search queries served.
-
4
Read cost and storage
One-off and monthly cost, and the memory to size your vector database for.
Frequently asked questions
text-embedding-3-small is priced at a few cents per million tokens, so embedding tens of thousands of documents usually costs less than a few dollars. Storage and hosting typically cost more than the embeddings themselves.
Between 256 and 1,024 tokens is typical, with 10 to 20 percent overlap. Smaller chunks give more precise matches; larger ones keep more context around each match.
Dimensions multiplied by bytes per value. A 1,536-dimension float32 vector is about 6 KB; the index and metadata add to that.
Yes. Half precision halves storage with almost no recall loss, int8 quarters it with a small loss, and binary vectors are tiny but usually need a re-ranking step.
Yes. Overlapping tokens are embedded in both neighbouring chunks, so heavy overlap raises both the embedding bill and the number of vectors stored.
Reviews of the Embedding Cost Calculator
What people who use the Embedding Cost Calculator think of it. Reviews come from visitors and are posted as written.
No ratings yet
Yours would be the first.
- 5 0 reviews rated 5 stars
- 4 0 reviews rated 4 stars
- 3 0 reviews rated 3 stars
- 2 0 reviews rated 2 stars
- 1 0 reviews rated 1 star
0 reviews
Newest first
No reviews yet
Be the first to rate the Embedding Cost Calculator. It takes half a minute and no account.
Write the first reviewMore AI tools
MCP Config Generator
Generate MCP server config for Claude Desktop, Claude Code, Cursor, VS Code and Windsurf.
AI Image Prompt Generator
Build detailed image prompts for Midjourney, Stable Diffusion and DALL-E.
ChatGPT Export to Markdown
Turn conversations.json from a ChatGPT or Claude export into readable Markdown files.
Token Counter
Estimate how many tokens your text costs before you send it.
Few-Shot Example Builder
Turn input and output pairs into a clean, consistent few-shot prompt block.
Midjourney Parameter Builder
Add aspect ratio, stylize, chaos, style references and more to a Midjourney prompt.
Fine-Tuning Dataset Formatter
Turn question and answer pairs into a JSONL training file for OpenAI, Llama or Gemini.