Google launches Gemini Decisions API to accelerate AI decision-making speed

·by Henderson·Engineering
Google launches Gemini Decisions API to accelerate AI decision-making speed
Key Points
  • The Gemini Decisions API is designed for rapid predefined decision-making.
  • Developers can submit various questions and handle multiple types of data simultaneously.
  • The model outperforms competing systems in benchmark tests.
  • Google's technology enhances the speed and efficiency of automated systems.

Google has launched the Gemini Decisions API, a new endpoint powered by its Gemini Decision 1 model, aimed at helping AI agents quickly make predefined decisions without generating lengthy text responses. The interface allows developers to evaluate up to 1,000 independent questions simultaneously within shared multimodal inputs that include text, images, audio, video, and PDFs. Gemini Decision 1 does not generate responses verbatim but scores predefined options in a single forward pass, returning probabilities and confidence scores with an end-to-end latency of 150 to 250 milliseconds.

Google states that the model is designed for limited tasks, such as selecting tools, checking policy compliance, routing queries to the correct model, and performing background scoring.

Developers can submit yes or no checks, multiple-choice questions with up to 255 options, or scoring requests based on predefined criteria in a single API call. According to Google, the median computation time for the model provided by the API is between 50 and 100 milliseconds, with performance remaining nearly constant when processing 1 to 100 questions per request. The model supports a context window of up to 256,000 tokens and can handle multiple types of data simultaneously. This allows developers to combine inputs such as audio recordings, images, and documents without relying on separate processing systems.

Model Performance Compared to Competitors

Google also claims that the model outperforms competing systems in benchmark tests covering agent routing, reasoning, ranking, and multimodal understanding, while achieving the lowest calibration error. The company states that the cost of Gemini Decision 1 is nearly one-fifth that of Gemini 3.5 Flash-Lite, using its shared benchmark suite's Interactions API. This launch comes amid increasing competition for specialized AI decision-making tools, with OpenAI launching its Decisions API powered by its GPT-6 Luna model in October.

The company claims that the service can deliver decisions up to 10 times faster and supports up to 400 questions per request. Similar to Google's product, the API offers probability-based checks, multiple-choice options, and standard-based scoring, with pricing starting at US$0.10 per million input tokens (approximately HK$1).

The cost of the Decisions API is US$0.10 per million input tokens (approximately HK$1), with no charges for output tokens. Developers only need to pay a one-time fee for shared input data, regardless of how many questions are evaluated, although each question will increase the number of instruction and standard tokens. Google has already deployed this technology across several of its products, demonstrating that specialized decision models can enhance the speed and efficiency of automated systems. In Gemini Notebook, the API routes queries from various sources to the appropriate model tier and reasoning level.

According to Google, this approach can reduce the latency of everyday queries by up to 40%, while reserving more powerful reasoning models for complex tasks.

Various Application Scenarios and Integrations

The technology also drives rapid design experiments in Google's Stitch platform, enabling users to generate interactive product wireframes and make changes via voice commands or touch input. Meanwhile, Google's CC Assistant utilizes the API to select emoji reactions from over 200 options in an instant. Google Search is also using this technology as a safety layer to detect adversarial behavior involving Model Context Protocol (MCP) tools. In addition to these applications, the API can help developers build systems that automatically select appropriate tools, assess incoming information, and determine when human intervention is needed.

The Decisions API is now available through Google AI Studio, including a free prototyping tier. Google is also launching integration services for developer platforms such as OpenRouter, LangChain, Vercel, LiteLLM, and Pydantic AI.

Market Competitiveness of Gemini Decisions API

As demand for AI decision-making tools grows, Google's Gemini Decisions API showcases its competitive advantage in the market with its high performance and multimodal processing capabilities. This API not only processes a large number of questions quickly but also enhances the efficiency of automated systems across various applications, meaning developers can apply it more flexibly in different scenarios. Additionally, the cost-effectiveness of this technology is likely to attract more developers, further expanding its market influence.

H
About the author
Henderson