# Token Pricing

> What you pay to use a model. There is a set price for each million tokens.

**Field:** Software Development  
**Category:** AI Engineering  
**Also known as:** LLM Pricing, Price per Million Tokens, API Pricing (LLM)

## Think of it like

Think of it like paying for sweets by weight. A bigger bag costs more, and some sweets cost more per bag.

## Definition

Token pricing is how model makers charge for API use. You pay a set price for each million tokens. Input and output have their own prices, and output most often costs four or five times more. Big, smart models cost far more per token than small, fast ones. Some also charge less for input that was cached or for batch jobs. To find the cost of a call, add the input tokens times the input price to the output tokens times the output price.

## Example

Lucy compares two models for tagging support tickets. The small one costs a tenth as much per token and tags just as well, so she picks it.

## Why it matters

It turns a choice, like which model to use or how much chat to send, into a cost you can plan.

## Common mistake

Picking the biggest model for every task. Many simple jobs run just as well on a small model at a fraction of the price.

## Three principles

- **Per million tokens.** The usual unit.
- **Output costs more.** Often several times.
- **Model size matters.** Small is cheap.

## Related terms

- [Token Usage](https://learn.benibix.com/term/development/token-usage)
- [LLM Batch API](https://learn.benibix.com/term/development/batch-api)
- [Prompt Caching](https://learn.benibix.com/term/ai/prompt-caching)
- [Model Fallback](https://learn.benibix.com/term/development/model-fallback)
- [Model as a Service (MaaS)](https://learn.benibix.com/term/ai/model-as-a-service)

![Two price tags per million tokens, a small one for input and a bigger one for output. A small model's tags are far cheaper than a large model's.](https://learn.benibix.com/diagrams/dev-token-pricing.svg)

---

Source: https://learn.benibix.com/term/development/token-pricing
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-20
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
