Purpose
Veo is Google DeepMind's video generation model. Official materials indicate that it has been replaced in the Gemini app by the multimodal model Gemini Omni (previously the Veo 3.1 model). Gemini Omni is positioned as a multimodal AI video generation and editing model that helps users turn ideas into videos, acting as a creative partner across the creative process from concept to finished clip.
Main capabilities
- Combine multiple types of materials such as text, images, and videos to generate videos.
- Generate videos based on reference images, supporting up to 5 reference images.
- Extend scenes and continue unfolding the story.
- Edit videos through natural conversation, enabling operations such as background replacement, clothing changes, style transfer, character replacement or substitution, lighting adjustment, video stabilization, and background changes while preserving details.
- Try curated style templates with one click to explore different possibilities.
- Use prompts to have Gemini suggest plot directions, expand the story, and lengthen the video.
- Add AI avatars to create content using an appearance and voice similar to your own, without needing to upload an image each time.
- Remix an existing portfolio (gallery).
How to use
In the Gemini app, users can create videos as if having a conversation: they can start from scratch, remix a portfolio, or try ready-made templates. When editing, simply state in the conversation what you want changed. Using an AI avatar is an optional feature, and only the user who created that avatar can use it for video creation; the traditional method of uploading your own photo can also continue to be used. Gemini Omni 1.1 Flash is described as being able to understand the surrounding world, add animation effects to photos, or create videos from any input; it is based on Gemini's world understanding and native multimodal capabilities, produces output that conforms to real-world logic, and supports gradually shaping results through natural conversation, completing AI video editing with prompts alone.
Applicable scenarios
Suitable for users who need to create and edit videos from concept to clip, such as animating photos, generating videos from multimodal materials, editing existing videos while preserving details, iteratively adjusting visuals through conversation, and appearing on camera using avatars.
Limitations and requirements
- Requires a Google AI subscription; feature availability varies by subscription tier and region.
- Intended for users aged 18 and above.
- All videos generated in the Gemini app have an invisible SynthID watermark embedded to identify Google AI-generated content; an official verification tool is also provided, allowing users to upload images, videos, and audio to ask whether they were generated by Google AI, and Gemini will check SynthID and answer based on its own judgment.
- Specific pricing, the features corresponding to each subscription tier, the list of available regions, and technical specifications such as video duration and resolution are not yet provided in the materials.
