Google launches Gemini Decisions API, handling up to 1,000 independent questions per request

Google on November 11, 2026 introduced the Gemini Decisions API, a new endpoint that evaluates as many as 1,000 independent questions in a single request, according to a report from Unite.ai.

The endpoint is powered by Gemini Decision 1 and is available through the Gemini API in Google AI Studio.

What it does

The API accepts inputs across five formats: text, images, audio, video and PDFs. Google built the endpoint for bounded judgments rather than prose or explanations. The use cases named are:

• Tool selection
• Policy checks
• Model routing
• Context grading
• Ranking

A request combines one shared input with independent questions about it.

Key details

• Endpoint: Gemini Decisions API
• Underlying model: Gemini Decision 1
• Availability: Gemini API in Google AI Studio
• Questions per request: up to 1,000, each independent
• Input types: text, images, audio, video, PDFs
• Output focus: bounded judgments, not prose or explanations

What this means

The Gemini Decisions API is designed for classification-style work inside larger systems, such as routing a request to a model, choosing a tool, or checking content against a policy. Because one request can carry up to 1,000 independent questions about a shared input, a single call can cover many such judgments at once. Developers who use Gemini for routing or grading can now do so through a dedicated endpoint in Google AI Studio instead of prompting a general-purpose model for open-ended text.