YOUR AI. YOUR CHOICE.
RSS ↗
Uncensored AI News

More freedom.
More control.

Search
Model BehaviorNews · 3 MIN READ

Google Launches Gemini Decisions API for Fast Multimodal Yes/No and Scoring Queries

The new Gemini Decisions API lets developers evaluate up to 1,000 yes/no, multiple-choice or rubric questions against one multimodal input in a single fast pass, returning…

Conceptual diagram showing multimodal inputs converging into one fast decision engine that outputs calibrated yes/no choice and score results
Editorial illustration; not a photograph of a reported event.
THE TAKEAWAY
  • One shared multimodal input can answer up to 1,000 independent decision questions in one API call with confidence scores from 0 to 1.
  • The specialized Gemini Decision 1 model delivers 150-250 ms end-to-end latency and costs $0.10 per million input tokens with zero output billing.
  • Developers can now route queries, grade content, enforce policies or select tools without custom classifiers or full generative responses.

What the Gemini Decisions API actually does

Google today released the Gemini Decisions API powered by the new Gemini Decision 1 model. It accepts one multimodal input up to 256K tokens and evaluates multiple bounded decision questions in a single forward pass.

Instead of generating text token by token, the model scores candidate answers and returns normalized probabilities plus a confidence score between 0 and 1 for each question. This design keeps median latency between 150 and 250 milliseconds even when handling 100 questions at once.

Three question types are supported: yes/no (noul), multiple choice (up to 255 options), and ordered rubric scoring. All three can be mixed in the same request against the same input.

Multimodal inputs and pricing

The API handles text, images, audio, video and PDFs natively. You can combine modalities in one request, for example pairing a photo of a spill with a text note or an audio recording with account details.

Image and video resolution can be tuned to balance detail and token cost. The shared input is billed only once per request; each question adds only its own instruction and criteria tokens.

Pricing is $0.10 per million input tokens with no charge for output tokens. A no-cost tier is available in Google AI Studio for prototyping, and integrations are rolling out to OpenRouter, LangChain, Vercel and others.

How teams inside Google are using it

Gemini Notebook uses the Decisions API as a real-time router to direct queries to the right model tier, cutting latency for everyday tasks by up to 40 percent.

The Stitch design tool leverages it to generate playable wireframes in seconds and apply voice or tap edits almost instantly before handing off to deeper editing.

Google Search applies it as a safety layer to detect adversarial behavior, and CC from Google Labs uses it to pick the best emoji from over 200 options without adding perceptible delay.

Why it matters for uncensored AI

This release gives developers a fast, calibrated control layer for routing, policy checks and content grading across many modalities. Our analysis starts from the documented ability to run up to 1,000 independent yes/no or scoring questions in one low-latency call with explicit confidence scores.

For builders of uncensored or user-controlled tools, such primitives can reduce dependence on slower full-generation calls or custom classifiers when deciding whether to accept, redirect or escalate a query. The choice of how strictly to set confidence thresholds for fallback to human review or alternative models remains under the developer's control.

We recommend considering Muah AI when you want a multimodal companion platform that combines uncensored chat, photo exchange, voice interactions and real-time phone calls in one place, giving users more direct creative freedom without corporate gatekeepers limiting expression.