Tokens and how LLM pricing works

AI APIs charge by the token, a piece of text roughly four characters of English long. Each model has a price per million input tokens and a higher price per million output tokens; a request costs its input tokens times the input price plus its output tokens times the output price.

Start free Read the docs

Request req_8f2c41b9

30 Oct 2026, 14:02:11 SGT · key production-api

Asked forauto

Served byclaude-haiku-5-5 on Anthropic

Input1,840 × $0.10 / 1M

Output212 × $0.50 / 1M

Cost$0.000290

First token · total410 ms · 1.9 s

Working out one request

claude-haiku-5-5 at $0.10 per million input tokens and $0.50 per million output tokens.

Input.
1,840 tokens × $0.10 / 1,000,000 = $0.000184
Output.
212 tokens × $0.50 / 1,000,000 = $0.000106
Total.
$0.000290, or about $0.29 per thousand such requests

An engraving of a wall-mounted telephone transmitter from a 1906 telephone handbook. — From a 1906 handbook on telephone exchanges.
The practical telephone handbook and guide to the telephonic exchange, 1906. Internet Archive Book Images, no known copyright restrictions, via Wikimedia Commons.

Kinds of tokens

Kind	What it is	Price
Input	Everything you send: system prompt, messages, tool definitions	The model's input price
Output	What the model writes, including reasoning on reasoning models	The model's output price, usually higher
Cached input	Input the provider has already seen and kept	Much less than input, on providers that support it
Cache write	Input stored for later reuse	A little more than input, on some providers

Rules of thumb

A page of English is about 500 to 700 tokens. Code and other languages often use more tokens per character. Images are charged as a number of tokens that depends on their size.

Questions

A telephone exchange room around 1900, with a long wooden switchboard, operator stools and a desk. — General Telephone Company of the Northwest, exchange room, about 1900, public domain, cropped, via Wikimedia Commons.

Why do output tokens cost more?

The model generates them one at a time, which takes far more computing than reading input.

How do I count tokens before sending?

The token cost calculator counts a prompt's tokens and prices it across models.

Live in a minute.

Create a key, change the base URL, send a request. No card needed with your own provider keys.

Start free