Agentic video understanding in Gemini

Source: Google for Developers on YouTube By Google for Developers

Summary

Google Gemini now offers agentic video understanding, a new capability in Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite. This approach processes long videos more efficiently, reducing token costs and improving accuracy by focusing on relevant parts. Maarten Grootendorst from Google DeepMind explains how this method works and its benefits for developers working with extensive video content.

Why it matters

This development addresses a key challenge in processing long videos: high token consumption. By using an agentic approach, Gemini can intelligently identify and analyze crucial segments, making video understanding more cost-effective and accurate. This is particularly relevant for developers and researchers working with large video datasets. Competitors like OpenAI and Anthropic are also investing in multimodal capabilities, but Gemini's specific agentic method for video offers a distinct advantage in efficiency. Future developments will likely focus on further refining these agentic capabilities for broader applications.

Read this on Google for Developers on YouTube
Opens in a new tab. Subvolts summarizes and links; the full piece belongs to Google for Developers on YouTube.
Where the other five stand

Related: OpenAI: Introducing GPT-6 Astra: the most intelligent and aligned model in the world. · Anthropic: GPT‑6 Astra · Microsoft: Meet MAI-Transcribe-2: A faster and more accurate speech recognition model · Meta: MetaRoCE: A New RDMA Transport Built for AI-Scale Ethernet · xAI: Sam Altman "AGI by December"

Hype check
2/5Worth a look

Rated low: routine. Worth knowing, not worth rearranging your day for.

Who's talking about it
Prior coverage our earlier items on the same thing
Published
Source
Google for Developers on YouTube (youtube.com)
Author
Google for Developers
Company
Google · Official · Videos
People
Maarten Grootendorst
Products
Gemini 3.7 Flash, Gemini 3.6 Flash, Gemini 3.5 Flash-Lite
Summary by
Subvolts, using an AI model (how we work). Spotted a mistake? Tell us.

Questions people ask

What is agentic video understanding in Gemini?
It is a capability that processes long videos more efficiently by reducing token costs and improving accuracy on relevant parts of the video.
Which Gemini models support this feature?
Agentic video understanding is available in Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite.

More from Google for Developers on YouTube 14 more

Everything from Google for Developers on YouTube →

Page generated Sep 3, 2026. Summaries are Subvolts' own; the story belongs to Google for Developers on YouTube.