Today, we’re introducing the Gemini Decisions API, a new endpoint for fast, closed-set decisions powered by our new Gemini Decision 1 model (#gemini-decision-1), available through the Gemini API in Google AI Studio. The Decisions API gives you purpose-built primitives for bounded choices like tool selection, policy checks, model routing, and context grading. You can pass one #multimodal input across text, images, audio, video, and PDFs (up to 256K tokens) along with up to 1,000 independent questions. Rather than generating answer text token by token, Gemini Decision 1 scores your candidate options in a single forward pass, returning probabilities and confidence scores in 150-250 ms end-to-end latency at $0.10 per 1M input tokens with zero output tokens billed.
On benchmarks spanning #agent routing, reasoning, ranking, and multimodal understanding, Gemini Decision 1 leads across the suite, with the lowest calibration error. Gemini Decision 1 scores options without token-by-token decoding, keeping median latency 50-100 ms of model compute (150-250 ms end-to-end), across multimodal inputs and contexts up to 16K tokens, remaining nearly flat from 1 to 100 questions per request. With zero output tokens billed, Gemini Decision 1 with the Decisions API is nearly 5x cheaper across the shared benchmark suite than Gemini 3.5 Flash-Lite with the Interactions API. Built on Gemini’s native multimodal architecture, Gemini Decision 1 can process native multimodal evaluation across three decision types in one pass.
[AgentUpdate Depth Analysis]The launch of the Gemini Decisions API marks a pivotal shift in AI Agent infrastructure from “generative” to “decisional.” Traditional LLM APIs center on text generation, whereas the Decisions API abstracts closed-set decisions into primitives that directly output probability distributions, drastically reducing latency and cost in agent loops. Compared to OpenAI’s Structured Outputs or Anthropic’s Tool Use, Gemini Decision 1 handles up to 1,000 questions in a single forward pass and supports multimodal inputs, enabling near-real-time decision-making for complex agent workflows such as multi-tool routing and policy compliance checks. Its lowest calibration error means agents can more reliably rely on confidence scores for self-correction or escalation. From an ecosystem perspective, this could spawn a new class of “decision-layer” middleware, where agent frameworks like LangChain and CrewAI integrate the API to replace some LLM calls, optimizing cost and performance. In the long run, the divergence between decision APIs and generation APIs will push agent architectures toward greater modularity and efficiency, making real-time multimodal agents feasible.
