Sakhanda Wire
NVDA $153.26 -1.85% MSFT $506.01 -2.31% GOOGL $280.39 -1.17% META $620.55 -0.94% AMZN $257.19 -0.22%
← Back to the news

Google adds agentic video processing to Gemini Flash models

Google has introduced agentic video understanding across its Gemini Flash models. Instead of reading an entire video at a fixed rate of one frame per second, Gemini can decide which sections to inspect, how densely to sample them, and whether to use frames, audio or transcripts. The feature is intended to retrieve only the material needed for a prompt.

Google reports up to 88% fewer tokens, 66% lower costs and 7% higher accuracy on standard video benchmarks. The gains are focused on longer material, including how-to videos and multi-hour recordings. Static processing remains the default and is described as preferable for clips shorter than five minutes or tasks requiring frame-by-frame precision.

The capability is available as a hosted API feature through Google AI Studio and the Gemini Enterprise Agent Platform, with file uploads and public YouTube URLs supported. It cannot be self-hosted and has no open weights. Standard Gemini API token pricing applies. Agentic processing is enabled through a processing field and is supported on Gemini 3.8, 3.7, 3.6 Flash and 3.5 Flash-Lite. API responses expose processing calls and results, while reasoning and loaded media are counted separately as thought and tool-use tokens.

Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.

Originally published by MarkTechPost on

Read the original on MarkTechPost ↗

The reporting and the image belong to MarkTechPost and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news