Google Unveils Innovative Video Analysis Tool for Enhanced Accuracy at Lower Costs

Google Introduces New Video Analysis Feature This week, Google announced its new “agentic video understanding” capability for the Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite models. This feature allows users to analyze videos with greater accuracy without incurring additional costs.

Introduction of the New Video Analysis Feature

Google has introduced the “agentic video understanding” feature for its Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite models. This enhancement enables users to analyze videos more accurately without adding extra costs.

Reduction in Token Usage and Costs

With this new capability, token consumption can be reduced by up to 88%, and costs may decrease by as much as 66%, while the accuracy of analysis improves by up to 7%. These advantages are particularly evident when processing longer videos.

How to Use the New Feature

Users can take advantage of this feature by configuring their API settings to “agentic” within Google AI Studio or the Gemini Enterprise Agent Platform. This feature becomes available swiftly and easily.

Transforming Video Processing

The new capability significantly alters the way long videos are processed, allowing developers to view and analyze only the essential sections of the video. This functionality can be applied across various applications.

Support for YouTube Videos

This feature will soon be available to YouTube users, enabling them to receive high-quality responses based on video content.