AI Briefing
KOSign in

Google Gemini Decision 1 Available on AI Gateway for Multimodal Routing

·2026.11.11 09:00

Key point

The multimodal model processes text, images, audio, video, and PDFs to return Boolean probabilities, choices, and scores for policy checks and ranking.

Details

  • Google Gemini Decision 1 is now available on AI Gateway, supporting text, images, audio, video, and PDF inputs.
  • The model returns Boolean probabilities, choices, and scores to route requests, check policies, filter agent tools, or rank results without generating output tokens.
  • Access is available via google/gemini-decision-1 using the AI SDK, TypeSafe-compatible API, or Decisions API.
  • PDF, audio, and video inputs are supported through AI SDK 7, while TypeSafe-compatible and Decisions APIs support text and image inputs.
  • Decision calls appear in logs and count toward budgets alongside other model requests on AI Gateway.
googleai-gatewaymultimodal-aidecision-modelsgemini-decision-1
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

Google AI BlogGoogle introduces Gemini Decisions API for fast, calibrated multimodal decisionsVercel BlogGemini 4 Carbon available on AI Gateway with 50% discount through November 2026Vercel BlogVercel AI Gateway November Index: Decision models hit 26% of requests, Anthropic wins new teams 2:1Google AI BlogGoogle brings Gemini Live to Apple CarPlay for hands-free driving conversationsGeekNewsGoogle acquires Crusoe's Wyoming data center siteRedditAnt Ling Demonstrates Ling-3.2-flash Generating Illustrated Food Journals from Meal Photos