Google Introduces Gemini Video Understanding Feature to Reduce Token Usage

Google DeepMind has launched a video understanding feature for Gemini agents, available for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite models. The feature improves long-video analysis efficiency by dynamically scanning segments, reducing token consumption by up to 88% and analysis costs by 66%, while increasing accuracy by 7%. It is currently accessible via the Gemini API.