Google Launches Gemini Omni 1.1 Flash for AI-Assisted Video

Google introduces Gemini Omni 1.1 Flash, a professional production-oriented update. Some features are only accessible through Google’s professional tools, while the model ranks first in text-to-video according to Arena.ai but falls behind MiniMax-H3 for image-to-video.

Differentiated Access and Ranking Position

Among the new features of Gemini Omni 1.1 Flash, only the scene extension function is available in the Gemini app for Google AI Plus, Pro, and Ultra subscribers. Other features, such as keyframe transitions or advanced upscaling, require the use of Google’s professional tools, such as Google AI Studio or Flow. According to the independent ranking by Arena.ai, Gemini Omni 1.1 Flash holds the top position for generating videos from text, but it lags behind MiniMax-H3 for generation from images. Google presents this model as capable of producing "studio-quality" videos.

What Omni 1.1 Actually Enables

The scene extension function allows the AI to analyze a 10-second sequence and generate up to 40 additional seconds in 10-second segments. By defining a starting image and an ending image, the model can create a continuous video between these two keyframes, facilitating the creation of complex camera movements or zoom transitions. For rendering, the AI offers upscaling in 1080p or 4K, while quick previews in 360p are available to speed up prototyping.

Intended Uses and Launch Context

These technologies could enhance the quality of productions made with limited budgets. For example, Netflix is already using artificial intelligence to generate certain scenes that are impossible to produce within the allocated budget. The promise of "studio quality" put forth by Google could also inspire creativity among content creators on platforms like YouTube. Gemini Omni 1.1 Flash is an improved version of the original Omni Flash introduced in May, featuring new tools designed for production and professional use.