Google has released a Gemini Decisions API, backed by a model called gemini-decision-1, through the Gemini API and AI Studio. It returns probabilities and confidence scores for yes/no, multiple choice or rubric questions instead of generated text, with 150 to 250 ms end-to-end latency. Input is billed at $0.10 per million tokens with no output charge. Google says support in Gemini in Google Cloud is coming soon, and lists uses such as routing, policy checks and grading steps.
