Model Serving & Economics
Cost per million tokens
A common pricing unit for language-model API usage.
Example
A provider lists a price for processing one million input or output tokens.
Why it matters
It makes tiny per-piece charges easier to read and estimate at a larger scale.
You'll hear it
“Are those prices for the text we send or the answer?”
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
- Must I buy a million tokens at once?
- Usually no. The unit describes the rate; services commonly bill for the amount actually used.
- Can I compare two models using this price alone?
- No. Differences in reply length, repeated attempts and discounts can change the final cost.
- Does this tell me the price of a subscription?
- Not necessarily. Subscription plans may use message limits or other pricing arrangements instead.