Gemini 3.8 Flash keeps token prices flat, not total costs
The catch in Google’s cheaper workhorse pitch is hiding in the meter: Gemini 3.8 Flash takes extra reasoning steps and makes iterative tool calls on difficult tasks, so it can consume more tokens even at the same introductory rates, $0.75 per million input and $3.75 per million output. That tradeoff matters for agents and long-running coding jobs, where better answers can quietly become larger bills. Developers can lower effort levels when compute efficiency matters.