AI
Google Launches Gemini Decisions API for Multimodal Judgments in a Single Forward Pass
Google has released the Gemini Decisions API, powered by the Gemini Decision 1 model, for closed-option tasks such as tool selection, policy checks, model routing, and context scoring. It can process text, images, audio, video, and PDFs, ask up to 1,000 independent questions in one request, and return probabilities and confidence scores.
Google AI announced the launch of the Gemini Decisions API, available through the Gemini API in Google AI Studio. The interface is designed for decision tasks with a clear set of candidates, including tool selection, policy checks, model routing, and context grading, rather than generating open-ended responses word by word.
Gemini Decision 1 can score candidate options in a single forward pass. The interface supports multimodal inputs such as text, images, audio, video, and PDFs, with context of up to 256K tokens, and can ask up to 1,000 independent questions simultaneously on the same input. Google says end-to-end latency is 150 to 250 milliseconds, pricing is $0.10 per million input tokens, and there is no charge based on output tokens.
The API supports three types of questions: yes-or-no questions that determine whether a condition holds, multiple-choice questions with 2 to 255 options, and scoring questions that provide a continuous weighted score and probabilities for each level according to an ordered rubric. Each answer comes with a confidence score from 0 to 1, making it easier for applications to send ambiguous cases to fallback workflows or human review. Google says the model leads benchmarks in agent routing, reasoning, ranking, and multimodal understanding and has the lowest calibration error; on a shared benchmark, its cost was nearly one-fifth that of Gemini 3.5 Flash-Lite using the Interactions API.