Google has launched the Gemini Decisions API, a new endpoint powered by its Gemini Decision 1 model that answers bounded questions instead of generating text, Google DeepMind announced on 11 November 2026.

A single request can carry one multimodal input, covering text, images, audio, video and PDFs of up to 256,000 tokens, together with as many as 1,000 independent questions. Rather than producing an answer token by token, the model scores the candidate options in a single forward pass and returns probabilities and confidence scores.

Three kinds of question

The API supports three question types in one request. Yes-or-no checks estimate the probability that a stated condition holds. Multiple-choice questions select from between two and 255 options, returning the winner and normalised probabilities for every candidate. Rubric scoring grades an input across an ordered scale, returning a continuous weighted score and per-level probabilities. Every answer includes a confidence score from zero to one, so code can route ambiguous cases to a fallback or a human reviewer.

Speed and price

Google reports median model compute of 50 to 100 milliseconds and end-to-end latency of 150 to 250 milliseconds, across multimodal inputs and contexts up to 16,000 tokens. Gemini Decision 1 is priced at $0.10 per one million input tokens, with no output tokens billed, and a no-cost tier is available in Google AI Studio for prototyping. Google says the API is nearly five times cheaper across its shared benchmark suite than Gemini 3.5 Flash-Lite with the Interactions API, a comparison published by Google and not independently checked.

Already in use inside Google

The API is already running inside Google. Gemini Notebook uses it as a real-time router, directing queries to the best model tier and thinking level, which Google says delivers up to 40 percent lower latency for everyday queries. It is also used in the Stitch prototyping tool, in emoji selection in the CC app from Google Labs, and as a safety layer for tool use in Google Search. Ecosystem integrations are rolling out across OpenRouter, LangChain, Vercel, LiteLLM, Pydantic AI, Pipecat, LiveKit and OpenCode, with support in Gemini in Google Cloud listed as coming soon. Initial access runs through the Gemini Developer API v1beta.