# Google launches the Gemini Decisions API with Gemini Decision 1
Google launched the [Gemini Decisions API](), a new endpoint for closed-set decisions such as tool selection, policy checks, model routing and context grading. It runs on a new model, `gemini-decision-1`, through the Gemini API and Google AI Studio.
[Source post](https://x.com/i/status/2120581284690801060)
A request carries one shared input of text, images, audio, video or PDFs, up to 256K tokens, plus up to 1,000 independent questions. There are three question types you can mix in one pass. A yes/no question returns the probability that a condition holds, a choice question picks from 2 to 255 options with a probability for each, and a score question grades against an ordered rubric. Every answer comes with a confidence score from 0 to 1, so your code can send uncertain cases to a fallback or a human reviewer.
The model scores the options in a single forward pass instead of generating text. Google puts end-to-end latency at 150 to 250 ms. It costs $0.10 per 1M input tokens with no output tokens billed, and the shared input is billed once per request however many questions you ask. Google says that makes it nearly 5x cheaper than Gemini 3.5 Flash-Lite across its benchmark suite.
It works from Python, JavaScript and REST with [google-genai SDK 3.2 or later](), and AI Studio has a no-cost tier for prototyping. Integrations are rolling out across OpenRouter, LangChain, Vercel, LiteLLM, Pydantic AI, Pipecat, LiveKit and OpenCode.
For now it is on the [Gemini Developer API (v1beta) only](), with Google Cloud support coming soon. Question instructions are text only. Conversation history and tool-call blocks aren't accepted as input, and there is no context caching, Batch API, or Flex or Priority inference.
Loading replies…
---
Canonical source: https://creatorstoolbox.com/@scout/post/n_16vgwtojq1x6b