Google officially announced Gemini Omni, a next-generation multimodal artificial intelligence model designed for advanced video creation and editing. This model allows users to manipulate and generate video content by seamlessly combining inputs across text, images, video clips, and audio.
Gemini Omni Flash, the debut model in the new lineup, is currently rolling out in the Gemini app, Google Flow, and YouTube Shorts.
The core feature of the Gemini Omni architecture is a conversational editing engine. According to Google, this model allows creators to change their videos using a series of natural language prompts. The system preserves the context of previous instructions and ensures structural continuity across scenes, recurring characters, and complex visual elements throughout the editing timeline.
The model also supports highly flexible generative inputs, allowing users to direct video production using:
- Text prompts and structured scripts
- Images, hand-drawn sketches, illustrations
- Existing video clips as style or structure references
- Direct dictation (compatible with additional dictation formats planned for future releases)
- By blending Gemini’s advanced reasoning capabilities with dedicated video generation capabilities, the model generates scenes based on real-world concepts, including physics, historical accuracy, and durable visual consistency.
In addition to the core model, Google introduced a personalized avatar feature. This tool allows creators to generate synthetic video content by utilizing highly accurate digital versions of themselves, synthesized together with clones of their own voices.
To address safety and reliability concerns regarding AI-generated media, Google has confirmed that it will be automatically embedded in all video content created via Gemini Omni. Synth ID Digital watermark. These imperceptible metadata markers keep the generated content verifiable. Viewers and platforms can directly see the origin of Omni-generated videos through the Gemini app, Gemini on Chrome, and Google Search.
This announcement marks a significant expansion of Google’s creative AI ecosystem, which has previously focused on dedicated image generation tools and static editing tools like Nano Banana.
