Input tokens

Also called: input token, prompt tokens, tokens in, input cost, input pricing

The tokens an AI product sends to the model on each request: the instructions, the conversation so far, files and tool results. Providers charge less for them than for output, but they add up, because the whole conversation is usually sent again with every request.

For example: A chat that has run for an hour sends its long history along with every new question, so most of the bill is input tokens rather than answers.

Connects to 9

Token Concept
Output tokens Concept
Cache tokens Concept
Prefix cache Concept
Context window Concept
Model provider request Concept
Usage-based pricing Concept
Stateless Concept
Tool result Concept