Token burn refers to the rate at which AI systems consume tokens during input and output processing. Tokens are the units AI models use to read and generate text. A token may be a word, part of a word, punctuation, or a symbol. When you send a prompt to an AI model, the system processes input tokens and then creates output tokens in response.

If you’ve spent any time learning about AI, you’ve probably come across the word “token.” It shows up often in discussions about models, pricing, and efficiency. But what does it actually mean and why should it matter to you?