Google has officially launched the Gemini Decisions API, a new endpoint designed to streamline tasks requiring fixed-menu responses rather than open-ended generation. The tool is powered by the specialized gemini-decision-1 model and is currently accessible via Google AI Studio.
The API supports multimodal inputs, including text, images, audio, video, and PDFs, with a capacity of up to 256K tokens. Developers can submit a single input and perform up to 1,000 independent queries simultaneously, which is intended to improve processing speed for tasks like policy checks and tool selection.
Google positions this release as a solution for AI agent development and compliance workflows. By offloading micro-decisions to this dedicated endpoint, companies may be able to optimize the use of larger generative models, although specific pricing details for the service have not yet been disclosed.
The competitive landscape for decision-focused AI models has intensified rapidly. OpenAI introduced its own Decisions API in public beta on October 6, 2026, followed by Microsoft, which released its Microsoft-Decision-1 model on October 9, 2026. Google entered this niche on November 11, 2026.
