Model Serving & Economics
Output tokens
Tokens generated by a model in its response.
Example
Every word generated in the answer contributes to output-token usage.
Why it matters
Counting them helps teams track the expense and length of generated replies.
You'll hear it
“We only need a paragraph, not a full report.”
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
- Can I set a maximum reply length?
- Many services allow a limit, though a reply may stop abruptly if the limit is too small.
- Do hidden reasoning steps count toward usage?
- Some models charge for internal work as well as visible text. Check how the service reports usage.
- Does deleting an answer reverse its usage charge?
- Generally no. The processing has already happened even if you remove the reply from view.