AI API Cost Calculator

Before you pick a model for a product, it helps to know what it will cost at your real volume. This calculator takes the number of input and output tokens you expect per month and prices them for every model in our list, from the cheapest to the most expensive, using the standard rates on each provider’s official pricing page.

How to use the AI cost calculator

1. Enter the input tokens you expect per month: everything you send, including system prompts, chat history and retrieved documents.
2. Enter the output tokens per month: everything the model writes back.
3. The table sorts every model from cheapest to most expensive and shows input cost, output cost and the monthly total.
4. Check the notes under a model for long-prompt price tiers or scheduled price changes.

How AI API pricing works

Almost every provider bills per million tokens, with separate prices for input (what you send) and output (what the model generates). Output is usually several times more expensive than input, so a chatbot that writes long answers costs more than a classifier that returns one word, even with the same prompts.

The formula is simple: input tokens ÷ 1,000,000 × input price + output tokens ÷ 1,000,000 × output price. To estimate monthly tokens, multiply the tokens in a typical request (you can measure them with the token counter) by the number of requests per month. Remember that chat apps resend the whole history on every turn, so input grows with conversation length.

Real bills can be lower or higher than this estimate. Prompt caching and batch APIs give large discounts on repeated or non-urgent work; some models charge more per token once a single prompt passes a size threshold; reasoning models also bill their hidden reasoning as output tokens. Prices on this page come from each provider’s official pricing page, and the date they were last checked is shown under the table.

Frequently asked questions

How accurate is the estimate?

The arithmetic is exact for the standard per-token prices shown. The uncertainty is in your inputs: real traffic, conversation length and how much the model writes. Start with measured token counts from a few real requests and multiply by your expected volume.

Why is output more expensive than input?

Generating text is the expensive part: the model produces output one token at a time, while input can be processed in parallel. That is why providers charge several times more per output token, and why limiting response length is one of the easiest ways to cut costs.

Are prompt caching and batch discounts included?

No. The table uses standard prices. Caching repeated prefixes (such as a long system prompt) and sending non-urgent jobs through a batch API can reduce the bill substantially. Check the provider’s pricing page for current discounts.

What does “Prompts over 272K tokens” mean under some models?

Some providers price a request differently when its prompt is very long. If a single request’s input is above that threshold, the higher rates apply to that whole request. A monthly total can’t tell how big each request is, so the table uses the base price and lists the higher tier as a note.

How up to date are the prices?

They come from the providers’ official pricing pages and the check date is shown below the table. Prices change often, so confirm on the provider’s site before committing to a budget.