The Gemini Decisions API is a Google endpoint that answers closed questions, such as yes or no, which option fits or which rubric level applies, instead of writing free text.
In plain words
It runs on Gemini Decision 1, a model that scores the possible answers in one pass rather than producing words one at a time. It returns probabilities and a confidence score, and no output tokens are charged. One request can carry a shared input of text, images, audio, video or PDFs plus up to 1,000 separate questions. Google suggests it for routing requests, checking policy rules and choosing tools. Developers use it via Google AI Studio and the Gemini API.
A quick example
Imagine an insurance app that receives a scanned claim form and photos of a damaged car. In one request it asks whether the form is complete, which damage category fits and how severe the damage looks, then sends the claim to the right team based on the answers.
Where you'll meet it
- Routing each request to the right model or team.
- Checking content against many policy rules at once.
- Picking which tool an agent should call next.
Related
Sources
- DocsIntroducing the Gemini Decisions API, Google · read 11 Nov 2026
- DocsGemini Decision, Gemini API docs · read 12 Nov 2026
Checked by WhatIsAI against the sources above · 12 Nov 2026. Spotted an error? Report it.