01 / The Model
Output tokens
What the model writes back. Priced higher than input, because generating text takes more computation than reading it. Long answers cost real money.
How you’ll hear it
“Can we make it cheaper?”
“Ask for shorter answers. Output is the expensive part.”
Avoid
Asking for long, exhaustive answers by default.
The longer version
Output tokens are what the model writes back. They cost more than input because producing text takes more computation than reading it. Asking for concise answers or a fixed format is one of the simplest ways to cut both cost and waiting time.