Gemini Live Agent Challenge

Overview

This hackathon challenges you to build next-generation AI agents that go beyond basic text conversations, using multimodal inputs and outputs. You'll leverage Google's Live API and creative tools to solve problems or create new user experiences. The goal is to redefine interaction by creating agents that can see, hear, speak, and create.

Requirements

Participants need to build a new AI agent that uses multimodal inputs and outputs and must leverage a Gemini model. Agents should be built using either the Google GenAI SDK or ADK and must utilize at least one Google Cloud service. All projects need to be hosted on Google Cloud. You'll need to submit a text description of your project, a link to your public code repository with setup instructions, proof of Google Cloud deployment, an architecture diagram, and a demo video under 4 minutes. For bonus points, you can publish content about your project, automate your cloud deployment with scripts, or provide a link to your public GDG profile.

Prizes

There's a total of $80,000 in cash prizes. The grand prize is $25,000 cash plus Google Cloud credits, mentorship, and tickets to Google Cloud Next 2026. There are also category prizes of $10,000 each for the Best of Live Agents, Best of Creative Storytellers, and Best of UI Navigators. Additional prizes include $5,000 for Best Multimodal Integration, Best Technical Execution, and Best Innovation, along with $2,000 cash for five Honorable Mentions. Many prize winners also receive Google Cloud credits and tickets to Google Cloud Next.