Google announced the Gemini Decisions API, powered by the Gemini Decision 1 model, as a purpose-built endpoint for fast, closed-set decision making across text, images, audio, video and PDFs. The company said the API can evaluate up to 1,000 independent yes/no, choice and rubric-scoring questions in a single forward pass over a shared multimodal input.
Operation and decision types
According to Google, the model scores candidate options without token-by-token decoding, returning probabilities and confidence scores for each question. The API supports three question types: noul (Yes/No) with probability estimates, choice (Multiple Choice) selecting among 2 to 255 choices with normalized probabilities, and score (Ordered Rubric) which returns a continuous weighted score plus per-level probabilities. Every answer includes a confidence value between 0 and 1 to help route ambiguous cases to fallbacks or human review.
Google described the input as a single multimodal Content block that can combine modalities (for example, a call recording with account metadata or a scanned PDF with photos). The announcement notes the shared input can be up to 256K tokens, and Google reported performance across multimodal contexts up to 16K tokens with median model compute of 50–100 ms and end-to-end latency of 150–250 ms.
Performance, cost and integrations
Google published benchmark claims that Gemini Decision 1 leads across agent routing, reasoning, ranking and multimodal understanding, and that it produces lower calibration error than other models on the suite. The company also said the Decisions API is nearly five times cheaper across the shared benchmark suite than Gemini 3.5 Flash-Lite with the Interactions API, since output tokens are not billed.
Pricing for Gemini Decision 1 is $0.10 per 1M input tokens, with no charge for output tokens. Google noted the shared input is billed once per request regardless of question count, and that an API no-cost tier is available in Google AI Studio for prototyping. The company highlighted internal use cases such as dynamic model routing in Gemini Notebook, rapid canvas features in Stitch, emoji selection in CC from Google Labs and a safety layer for MCP tools in Search. Ecosystem integrations noted include OpenRouter, LangChain, Vercel and others.
Developers can obtain an API key in Google AI Studio and consult the Decisions API documentation for Python, JavaScript and REST guides.
Original source: Google AI