Google has introduced the updated Gemini 1.2 Flash model

Google DeepMind has announced the release of Gemini 1.2 Flash, an updated version of its high-performance model optimized for handling long video streams and reducing latency. The release aims to improve the efficiency of real-time multimodal data processing.

A key improvement is the support for an expanded context window, which allows for the analysis of video files up to 60 minutes long in a single request. Developers have optimized the model's architecture, resulting in reduced response times while maintaining high accuracy. The update also includes improved mechanisms for processing audio and visual signals, making the model better suited for creating interactive agents.

An important addition is native support for generating and editing videos up to 15 seconds long, as well as increased visual consistency in results. Furthermore, the cost of token processing has been reduced by 30%, making the model more accessible for scalable enterprise solutions. The model is already available for integration via API, providing developers with tools to build applications that require deep video content analysis without the need to split files into fragments.

For practitioners, this means the ability to integrate features for analyzing long video lectures, meeting recordings, or complex visual instructions directly into products. The increased context length and cost optimization allow for reduced infrastructure load when working with heavy multimedia data.