Unchanged tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition → prices do not imply unchanged task costs. In Artificial Analysis' Gemini 3.8 Flash evaluation, Google DeepMind's model produced more output and took more turns on agentic tasks than its predecessor. The resulting cost per task rose even while the announced introductory input and output rates stayed the same.
Reasoning effort changed the tradeoff again: lower settings reduced the measured cost and duration alongside the score. The report is therefore most useful as a comparison of specific configurations on the evaluator's task mix, rather than evidence that every request becomes cheaper or that a single model name implies one fixed level of performance.
Terms in this piece · Glossary
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
