---
Title: "What Are Tokens in AI? LLM Tokens Explained"
Url: "https://computer.io/glossary/tokens-in-ai"
Topic: "AI foundations"
Description: "Tokens are the small pieces of text, like words or parts of words, that AI language models read and generate. How tokens work and why they affect cost."
---

# What are tokens in AI?

Tokens are the small chunks of text, such as words, parts of words or punctuation, that language models read and generate. In English, one token is roughly four characters or three-quarters of a word. Tokens determine how much a model can process and how much it costs.

## Example

"Resetting passwords" might be split into tokens like "Reset", "ting", " passwords".

## Why tokens matter

| Area | Effect |
| --- | --- |
| Cost | APIs charge per input and output token |
| Limits | Context window is measured in tokens |
| Speed | More tokens take longer |

## Reducing token use

Retrieve only relevant context, summarize history and keep prompts concise.

## FAQ

### How many words is 1,000 tokens?

Roughly 750 English words.

### Do all models count tokens the same way?

No, each model has its own tokenizer.

### Are output tokens more expensive?

Often, depending on the provider.

## Related terms

- [Context window](https://computer.io/glossary/context-window)
- [Large language model (LLM)](https://computer.io/glossary/large-language-model)
- [How do you reduce LLM token costs for enterprise AI?](https://computer.io/answers/how-to-reduce-llm-token-costs)

## How Computer helps

- [Pricing](https://computer.io/pricing)

## Further reading

- [History of ai part 2](https://devrev.ai/blog/history-of-ai-part-2)