Google launches Gemini Decisions API and Gemini Decision 1
Original title: Gemini Decisions API: Fast, calibrated multimodal decisions
One request can score up to 1,000 closed-set questions for $0.10 per million input tokens, with output tokens free.
Google DeepMind introduced the Gemini Decisions API, powered by the new Gemini Decision 1 model (gemini-decision-1), in the Gemini API and Google AI Studio. It scores yes/no, multiple-choice, and rubric questions in one forward pass over text, images, audio, video, and PDFs, returning probabilities and confidence instead of generated text. Google says end-to-end latency is 150-250 ms and that, with no output tokens billed, it is nearly 5x cheaper than Gemini 3.5 Flash-Lite on a shared benchmark suite. Internal uses include model routing in Gemini Notebook, wireframes in Stitch, emoji selection in CC, and an MCP-tool safety check in Search.
Unconfirmed: Benchmark leadership, calibration, and the nearly 5x cost comparison are Google's own figures.
- A shared input can mix modalities up to 256K tokens and up to 1,000 questions.
- Choice questions cover 2 to 255 options and return a probability for each.
- Google cites 50-100 ms of model compute and 150-250 ms end-to-end.
- Search uses the API to detect adversarial behavior against MCP tools inside its latency budget.
- A no-cost prototyping tier is available in Google AI Studio.
- Vendors
- Organizations
- Google DeepMind
- AI models
- Gemini Decision 1
From the original article
5 min read at blog.google
Meet the Gemini Decisions API, powered by our state-of-the-art model Gemini Decision for fast, calibrated multimodal decisions.
By Min Choi · blog.google
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.