Background
Google introduced the Gemini Decisions API on November 11, 2026. It is powered by a new model, Gemini Decision 1 (gemini-decision-1), and the API is available through the Gemini API in Google AI Studio.[1]
What the API does
A single request can evaluate up to 1,000 questions in one pass. The questions can be yes/no, multiple-choice or rubric-scoring questions. The input can be text, images, audio, video and PDFs, up to 256K tokens of input.[1]
Each answer comes back with a probability and a confidence score between 0 and 1. Google said the end-to-end latency is 150 to 250 milliseconds.[1]
Pricing
Gemini Decision 1 is priced at $0.10 per million input tokens, and no output tokens are billed. AI Studio offers a no-cost tier for prototyping.[1]
Uses inside Google
Google described two uses in its own products. Gemini Notebook uses the Decisions API as a model router, which Google said gives up to 40% lower latency for everyday queries. Google Search uses it as a safety layer to detect adversarial behavior in MCP tools.[1]
MCP (Model Context Protocol) is a standard for connecting AI models to outside tools. Adversarial behavior here means attempts to manipulate a system through its inputs.
Ecosystem
Integrations including OpenRouter, LangChain, Vercel, LiteLLM, Pydantic AI, Pipecat, LiveKit and OpenCode were rolling out. Support in Gemini in Google Cloud is to come later.[1]
Sources
- 1
Cite this page: Google launches the Gemini Decisions API. TellWorld, https://www.tellworld.net/news/gemini-decisions-api-2026. Updated 2026-11-15