Google Gemini's new agent-based video analysis cuts token usage by up to 88 percent
What changed
Google is adding agent-based video analysis to Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite. Instead of reviewing every frame at one fixed rate, the systems choose which parts of a video to examine and how much detail to use. Google says this can reduce usage by up to 88% and improve accuracy, particularly for videos lasting several hours.
What this means for you
Developers using these Gemini models may be able to analyze long videos more efficiently and accurately, potentially lowering usage costs. The metadata says the feature is being added, but does not specify when it will be available or its pricing.
