Google Launches Gemini Decisions API, Cutting Notebook Latency by Up to 40%

Google has released the Gemini Decisions API, which Gemini Notebook uses to choose models and thinking levels for incoming queries. Google says this reduced latency for everyday requests in Gemini Notebook by up to 40 percent, and other Google products such as Stitch and Search also use the API. Developers can access it through Google AI Studio, with paid usage starting at $0.10 per million input tokens.

Why it matters: Developers building on Gemini can now use a routing API that claims to cut latency and cost by matching each query to the right model and reasoning depth.