Updated

Gemini Decision 1

Google introduced Gemini Decision 1, a multimodal model for fast, closed-set choices through the Gemini Decisions API. It scores up to 1,000 yes/no, multiple-choice, or rubric questions over shared text, image, audio, video, or PDF input in a single pass.

By No Code MBA · Facts checked

Provider
Google
Announcement date
Status at last check
Generally available

What changed

Google introduced Gemini Decision 1 as a new model behind the Gemini Decisions API, rather than a token-by-token text generator. Google says it evaluates shared multimodal input against yes/no, choice, and rubric questions in one pass, returning probabilities and confidence scores.

Access and availability

Create an API key in Google AI Studio and use the Gemini Decisions API with model ID gemini-decision-1. Google says a no-cost API tier is available for prototyping.

Official access documentation

What you can build

Use it for bounded choices such as routing or policy checks, and route low-confidence or consequential decisions to a human review path.

Limitations

The model is designed for closed-set decisions and returns scores rather than generated prose. Google documents up to 1,000 questions and input up to 256K tokens; performance and latency claims are provider-reported.

Sources and release history