Question 730
A fine-tuned model release doubles the average output tokens, while request rates remain unchanged and the cost dashboard shows no change because it tracks only prompt tokens. Which usage counter is missing?
Explanation
Correct Answer
Output tokens contribute to cost and latency; request counts or input tokens alone omit that increased consumption.
Incorrect Answers
Branch inventory is unrelated to generation usage.
Staff counts do not measure billed output.
Manifest bytes are not produced tokens.