Introducing agentic video understanding with Gemini
Gemini can now inspect only the video moments it needs, cutting token use by up to 88%.
Google says agentic video understanding is available for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite through the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform. The model can dynamically search frames, audio, and transcripts instead of processing video at a fixed frame rate. Google claims costs fall by up to 66% and accuracy improves by up to 7% on tested video benchmarks. The feature has no added fee beyond standard Gemini API token pricing, and Google says it will come to the Gemini app and later YouTube’s “Ask YouTube” feature. Google DeepMind's note
Google says agentic video understanding is available for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite through the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform. The model can dynamically search frames, audio, and transcripts instead of processing video at a fixed frame rate. Google claims costs fall by up to 66% and accuracy improves by up to 7% on tested video benchmarks. The feature has no added fee beyond standard Gemini API token pricing, and Google says it will come to the Gemini app and later YouTube’s “Ask YouTube” feature. Google DeepMind's note
score 7