Google Launches Gemini Omni 1.1 Flash Video Generation AI with 4K Support
Google has released Gemini Omni 1.1 Flash, a production-ready video generation AI model that supports up to 4K resolution, scene extension, and reference-based generation. The model offers developers enhanced control over video creation, including the ability to specify start and end frames for smooth transitions. This update intensifies competition in the generative AI video space by offering developers a highly controllable and cost-effective tool. Features like 360p preview generation at one-third the cost of standard 720p allow for rapid prototyping and storyboarding before committing to high-resolution rendering. The model allows users to extend videos in 10-second increments up to a total of 40 seconds, and supports a 3-second reference video to maintain character and context consistency. Pricing ranges from $0.03 per second for 360p previews to $0.30 per second for 4K final outputs.
## BACKGROUND
Generative video AI has rapidly evolved, but developers often struggle with maintaining visual consistency across frames and managing high rendering costs. Keyframe-based generation and reference-based video models address these issues by allowing creators to guide the AI using specific starting/ending images or short reference clips to ensure characters and backgrounds remain stable.