Google adds agentic video analysis to Gemini Flash models, claiming 88% lower token use
- Agentic video understanding is now available in the Gemini API for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite, supporting uploaded videos and YouTube videos in Google AI Studio and the Gemini Enterprise Agent Platform.
- Google says the feature cuts token consumption by up to 88%, lowers video-analysis costs by up to 66%, and improves benchmark accuracy by up to 7% versus static processing.
- Instead of ingesting video at a fixed frame rate, Gemini dynamically searches and inspects selected video segments across frames, audio, and transcripts, choosing what to inspect and at what speed.
- Google says the system supports sub-second moment retrieval, multi-hour video search, anomaly detection through higher-FPS resampling, and counting repeated actions or objects.
- Developers enable the feature by setting API processing to "agentic"; it uses standard Gemini API token pricing with no separate feature fee.