LLM API Cost Calculator
Enter your token volumes and your provider's prices to see what a chatbot, assistant or automation will cost before you build it.
Estimated cost
| Scenario name | Small model (example) | Mid-size model (example) | Large model (example) |
|---|---|---|---|
| Per request | $0.00062 | $0.00310 | $0.02 |
| Per 1,000 requests | $0.62 | $3.10 | $15.50 |
| Per day | $0.12 | $0.62 | $3.10 |
| Per month | $3.72 | $18.60 | $93.00 |
| Per year | $44.64 | $223.20 | $1,116.00 |
- Tokens per month
- 11,400,000
- Lowest monthly cost
- $3.72
- Small model (example)
Estimate tokens from a sample text
Paste a typical prompt or answer. The estimate uses common rules of thumb (about 4 characters or 0.75 English words per token); real counts depend on the model's tokenizer and the language.
- Words
- 0
- Characters
- 0
- Estimated tokens
- ≈ 0
Runs entirely in your browser. No data is sent to a server.
How it works
- Enter the average input and output tokens per request. If you are unsure, paste a typical prompt or answer into the estimator at the bottom and transfer the result.
- Set how many requests you expect per day and on how many days per month.
- Replace the example prices with the current per-million-token prices from your provider's pricing page. You can rename each scenario.
- If your provider supports prompt caching, enter the share of input that is reused and the cached price.
- Read the cost per request, per day, per month and per year for each scenario in the results table.
FAQ
Are the default prices real?
No. They are round example values to show how the calculation works. Prices change often and differ by provider and model, so always enter the current figures from your provider's pricing page.
How is the cost per request calculated?
Input tokens times the input price plus output tokens times the output price, each divided by one million. Cached input tokens are charged at the cached price instead of the normal input price.
How many tokens is a word?
As a rough rule, one token is about four characters or three quarters of an English word. German, French and Spanish usually need more tokens for the same content. The exact number depends on the model's tokenizer.
Why are output tokens usually more expensive?
Generating text is more compute-intensive than reading it, so most providers charge a higher rate for output. Long answers can therefore dominate your bill even when prompts are long.
Does the calculator include taxes or subscription fees?
No. It estimates usage-based API charges only. Add VAT, platform fees, minimum commitments or tool subscriptions separately.
Related articles
- How to Choose an LLM for Your BusinessAI strategy8 min
- How to Reduce LLM API Costs: 12 Practical TacticsCosts & ROI8 min
- LLM API Pricing Explained: How Token Costs Add UpCosts & ROI8 min
- What Are Tokens in LLMs? A Plain-English ExplainerCosts & ROI7 min
More tools
- AI Automation ROI Calculator
Find out whether automating a recurring task with AI pays off, with review time and running costs included rather than ignored.
- AI Prompt Generator
Fill in a few fields and get a well-structured prompt you can paste into any AI assistant. Start from a template or a blank page.
- EU AI Act Risk Checker
Get a first orientation on where an AI use case sits under the EU AI Act: prohibited, high-risk, transparency duties or minimal risk.