Model Serving & Economics
Input tokens
Tokens sent into a model as instructions, context, or user content.
Example
The full document and instructions count toward the request's input tokens.
Why it matters
Counting them helps estimate charges and check whether a request will fit the model's limits.
You'll hear it
“Do we need to send the whole conversation again?”
Can you control it?
Set by the team
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Do hidden product instructions count too?
- They can. The application may send instructions that are not visible in your message box.
- Does an attached document always count in full?
- Not always. Some products send only selected passages; others process much more of the document.
- Can repeated text receive a discount?
- Some services offer discounts for reused material. Availability and billing details depend on the provider.