Google API Pricing

Google’s Gemini API has the widest price spread of the three providers — from Gemini 3.1 Flash-Lite at $0.25 per million input tokens up to the Gemini 4 Argon flagship — and the most promotional pricing. Argon launched at an introductory $2 / $10 that doubles to $4 / $20 on December 1, 2026, and the current Flash rates are promotional until the end of 2026. A Gemini price is therefore a price with a date attached, and budgeting on today’s number without checking its expiry is the most common Gemini budgeting mistake.

The second structural difference is Argon’s unified token budget: input and output share a single 1M-token pool, so an 800K-token prompt leaves only 200K tokens for the response, unlike providers that cap output separately. Output prices on Gemini also include the model’s thinking tokens, which means reasoning-heavy workloads pay for tokens you never see. Against that, Argon’s cached input discount is the industry’s most generous at 95% off, and its Batch/Flex tier halves rates for latency-tolerant work.

The table below uses Google’s official API list prices, each stamped with a verification date and its promotional expiry where one exists. Gemini consumer subscriptions and Vertex AI enterprise pricing are separate schedules and are not mixed in.