Google DeepMind has introduced agentic video understanding, a new capability for its Gemini models that enhances video analysis accuracy while significantly reducing token usage and costs. This feature dynamically scans video segments using native tools, improving performance on tasks like moment retrieval, anomaly detection, and counting. It is now available through the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform, offering up to 66% cost reduction and 88% token savings. AI
IMPACT Enhances video analysis efficiency and accuracy, potentially lowering costs for developers working with long-form video content.
RANK_REASON New capability for a frontier model family (Gemini) with performance claims. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- 3.5 Flash-Lite
- 3.6 Flash
- Gemini
- Gemini 3.7 Flash
- Gemini API
- Gemini Enterprise Agent Platform
- Google AI Studio
- Google DeepMind
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →