Embedding Cost Calculator
Calculate monthly and annual costs of generating text embeddings with OpenAI or Cohere models.
Did this tool work for you?
How to use this calculator
- 1
Enter the number of texts you need to embed each month (documents, product descriptions, messages, etc.).
- 2
Enter the average number of tokens per text. A typical paragraph is 100–300 tokens.
- 3
Select your embedding model and compare costs instantly.
Frequently asked questions
What formula does the embedding calculator use?
Total tokens = texts per month × average tokens per text. Monthly cost = total tokens ÷ 1,000 × the model’s rate per 1,000 tokens. Annual cost is monthly × 12, and cost per 1,000 texts is monthly cost ÷ texts × 1,000.
Which rates are built in?
Per 1,000 tokens: 0.00002 for text-embedding-3-small, 0.00013 for text-embedding-3-large, 0.0001 for ada-002 and 0.0001 for the Cohere embed-english entry. They are figures built into this tool, equal to $0.02, $0.13, $0.10 and $0.10 per million tokens. Providers revise prices, so check each pricing page.
Why do ada-002 and Cohere show identical costs?
Both entries carry the same rate, 0.0001 per 1,000 tokens, in the tool’s table, so any workload prices identically. The real provider prices may differ now or later, and the Cohere entry does not name a version, so verify before relying on it.
How much more does text-embedding-3-large cost than 3-small here?
Using the table’s rates, 0.00013 ÷ 0.00002 is 6.5 times. On 600,000,000 tokens a month the tool shows $78.0000 against $12.0000. The calculator compares price only and says nothing about retrieval quality.
Do I pay again to re-embed a document?
Yes. Every text sent to the embedding endpoint is billed, whether it is new or an updated version of an old one. Include edited documents in the texts-per-month count, and add query volume if you embed user queries at search time.
Does the estimate include vector database or storage costs?
No. Only embedding tokens are priced. Storage, indexing, query compute and bandwidth for the vector store are separate and depend on your database and vector size.
Why is the monthly cost shown with four decimals?
Embedding workloads are often only cents per month, so two decimals would round the default example of $0.4000 to the nearest cent and hide differences between models. Annual cost is shown with two decimals.
Is there any discount or free tier modelled?
No. The calculation is a flat rate per token with no batch discount, free allowance or volume pricing.
Embedding cost calculator: monthly and annual spend per model
The calculation
Total tokens per month = texts per month × average tokens per text. Monthly cost = total tokens ÷ 1,000 × the model’s rate per 1,000 tokens; annual cost is twelve times that. Cost per 1,000 texts is monthly cost ÷ texts × 1,000, a convenient unit for comparing pipelines of different sizes. The monthly figure is shown to four decimals because small corpora often cost only cents. There is no setup or per-request fee in the formula, only the token total multiplied by one rate. That simplicity means the result is only as good as the two volume inputs, so estimate them from a real sample of your content rather than from a round number.
Rates stored in the code
These are figures built into this tool. Embedding prices are revised from time to time and a model name can cover several versions, so confirm against each provider’s pricing page.
| Model entry | Per 1,000 tokens | Per 1M tokens |
|---|---|---|
| OpenAI text-embedding-3-small | $0.00002 | $0.02 |
| OpenAI text-embedding-3-large | $0.00013 | $0.13 |
| OpenAI ada-002 | $0.0001 | $0.10 |
| Cohere embed-english | $0.0001 | $0.10 |
Worked examples
Each result was produced by running the calculator.
| Run | Total tokens | Monthly | Annual | Per 1,000 texts |
|---|---|---|---|---|
| 100,000 texts × 200 tokens, 3-small (defaults) | 20,000,000 | $0.4000 | $4.80 | $0.0040 |
| 2,000,000 texts × 300 tokens, 3-small | 600,000,000 | $12.0000 | $144.00 | $0.0060 |
| 2,000,000 texts × 300 tokens, ada-002 | 600,000,000 | $60.0000 | $720.00 | $0.0300 |
| 2,000,000 texts × 300 tokens, 3-large | 600,000,000 | $78.0000 | $936.00 | $0.0390 |
| 500,000 texts × 400 tokens, ada-002 or Cohere | 200,000,000 | $20.0000 | $240.00 | $0.0400 |
Reading the comparison
On the identical 600-million-token workload, the 3-large entry costs $78.0000 against $12.0000 for 3-small, a ratio of 6.5 to 1 that follows from the 0.00013 and 0.00002 rates in the table. The ada-002 and Cohere entries share one rate, so the tool shows identical costs for them. Price is the only thing compared: output quality, vector size and storage footprint differ between models and affect the vector database bill, which this calculator does not estimate. If quality is the deciding factor, run a small retrieval test on your own documents with each candidate before comparing costs.
Choosing the token count
Average tokens per text should reflect what you actually send to the embedding endpoint after any splitting. Overlapping chunks embed the same words twice, so overlap raises the count. Re-embedding a document after an edit is billed again, so the monthly texts figure should include both new and updated items. A search query is also a text: if you embed queries on demand, add that volume. Sampling real chunk lengths gives a better average than a guess of a paragraph.
Learn more from an authoritative source:
OpenAI Platform DocsFine-Tuning Cost Calculator
Estimate the training cost and ongoing inference cost of fine-tuning an OpenAI language model.
AI Image Generation Cost Calculator
Calculate monthly costs for AI image generation with DALL-E 3, Stable Diffusion, or Midjourney.
RAG Pipeline Cost Calculator
Estimate the monthly cost of a Retrieval-Augmented Generation (RAG) pipeline including indexing, embeddings, and LLM inference.
GPU Cloud Cost Calculator
Compare GPU rental costs across AWS, GCP, Azure, Lambda Labs, and Vast.ai for ML training and inference workloads.
AI Chatbot ROI Calculator
Calculate the return on investment of deploying an AI chatbot for customer support, including payback period and annual ROI.
LLM Latency Calculator
Estimate expected LLM response latency — time-to-first-token and total response time — for planning real-time AI applications.
Results are estimates for informational purposes only and do not constitute professional financial, medical, legal, or technical advice. Read full disclaimer →