Building on an AI API is cheap to start and surprisingly easy to overspend on at scale. Costs depend on four things: the price per million input tokens, the price per million output tokens (usually several times higher), how many tokens each request uses, and how many requests you make. This free AI API cost calculator does the arithmetic so you can budget a chatbot, an internal tool, a document-processing pipeline or an agent before the invoice arrives. Enter the prices from your provider's current pricing page. We deliberately don't hard-code model prices, because they change often and differ by region, tier and contract, so you always calculate with real numbers. Then add your average input and output tokens per request and your daily request volume. You'll see the cost per request, per 1,000 requests, per day, per month and per year. Optional fields cover cached input (prompt-caching discounts for repeated system prompts or documents) and a batch discount for asynchronous jobs that many providers offer at around half price. A second scenario lets you compare two models or prices side by side. Not sure how many tokens your prompts use? Paste a sample into our AI Token Counter first. Everything is calculated in your browser.
Estimate API costs
Use current prices from your provider’s pricing page. Example values are placeholders, not real model prices.
How to estimate your bill
- Copy input and output prices per 1M tokens from your provider.
- Estimate tokens per request (system prompt + user message + context in; the reply out).
- Enter requests per day. Add caching or batch discounts if you use them.
Ways to cut costs
- Shorten system prompts and trim conversation history.
- Cache repeated context.
- Use a smaller model for simple steps.
- Batch non-urgent jobs.
FAQ
Why don’t you list model prices?
They change frequently and vary by tier and region. Entering current prices yourself keeps the estimate accurate.
What is a token?
A chunk of text, roughly ¾ of an English word on average. 1,000 tokens is about 750 words.
Why are output tokens more expensive?
Generating text costs the provider more compute than reading it.
Is anything sent anywhere?
No. It’s calculated in your browser.
Comments
Post a Comment