Google launches an API that quickly scores yes/no, multiple-choice and rubric questions

Google has launched the Gemini Decisions API, powered by a new model called Gemini Decision 1 (gemini-decision-1), for fast, closed-set decisions across text, images, audio, video and PDFs. The API can evaluate up to 1,000 independent questions in a single request against one shared multimodal input and returns calibrated probabilities and confidence scores to help route queries, grade content or check policy rules.

The model scores candidate options in a single forward pass rather than generating answer text token by token. Google says that means the model looks at the choices and the input once and assigns probabilities, instead of building an answer word by word, which keeps model compute median latency at about 50-100 ms and end-to-end latency at roughly 150-250 ms.

Some technical terms explained: input tokens are the units Google counts when it measures the size of the text or multimodal content you send (larger files use more tokens); a forward pass means the model evaluates all candidate answers at once instead of generating text step by step; and calibrated probabilities are confidence scores the model gives so your system can tell when an answer is uncertain and route it to a fallback or a human reviewer.

The Decisions API supports three decision types in a single call:

• noul (Yes/No): estimates the probability a yes/no condition holds, with optional true/false description criteria;
• choice (Multiple choice): selects from 2 to 255 options and returns normalized probabilities for every candidate;
• score (Ordered rubric): grades inputs on an ordered rubric (starting at 0), returning a continuous weighted score and per-level probabilities.

On price, Google lists Gemini Decision 1 at $0.10 per 1 million input tokens and charges $0 for output tokens. Because the shared multimodal input is billed only once per request regardless of how many questions are evaluated, Google says the Decisions API is nearly five times cheaper across its benchmark suite than using Gemini 3.5 Flash-Lite with the Interactions API.

Developers can access the API with an API key in Google AI Studio and follow Python, JavaScript and REST guides. Google also named ecosystem integrations including OpenRouter, LangChain, Vercel, LiteLLM, Pydantic AI, Pipecat, LiveKit, OpenCode and the google-genai SDKs, and demo apps such as Doorstep Parcels, Speed Sketch and Find the Zoo Keeper. Google says support is coming soon to Gemini in Google Cloud.

For organisations in Africa, two practical points matter. First, latency and legal rules depend on where the model runs. Google Cloud says regionally available products meet data residency standards; African teams should check whether Gemini Decision 1 is deployed in a nearby region before routing sensitive data abroad.

Second, Google Cloud documentation explains the trade-offs between latency, reliability and cost. Teams should review those options and choose the approach that fits their workload: low-latency interactive traffic has different needs from high-volume batch jobs. Those choices affect both response time and billing, so teams must weigh real-time Decisions API calls against batch processing for large asynchronous workloads.

Organisations can start building immediately by getting an API key in Google AI Studio and following the Decisions API documentation and SDK guides.

Newzroom dot Africa has reached out to Google and Google Cloud for comment and is awaiting a response. This story will be updated when they respond.

Source: Google