Google announced the launch of Gemini Omni, a new model designed to create content from a variety of inputs, with an initial focus on video. The first version, dubbed Gemini Omni Flash, is rolling out today to users of the Gemini app, Google Flow, and YouTube Shorts.
According to Google, Gemini Omni is considered “the next step” beyond its previous models, including Nano Banana and the existing video generator, Veo 3.1. The model enables users to combine images, audio, video, and text as input to generate high-quality videos that are grounded in advanced real-world knowledge.
Editing capabilities allow users to modify videos through natural conversation, building upon previous instructions for consistency in characters and elements. This contrasts with Veo 3.1, which was restricted to generating video content based solely on prompts and images.
Gemini Omni Flash allows users to shoot a video and then request modifications, transforming their initial content into something new. Google stated, “Your video becomes a starting point for something you never could have filmed yourself,” indicating that users can alter actions, add characters, and change settings seamlessly.
Video: Google
The model better understands physical principles such as gravity and kinetic energy to generate more realistic scenes. Gemini Omni integrates knowledge from various domains, including history, science, and cultural context, to enhance storytelling within the generated content.
The application can produce visual explainers from simple prompts to simplify complex ideas. However, initial audio features will support only voice references.
Gemini Omni also includes functionality to create a digital avatar based on the user’s appearance and voice. Google emphasized that it has established “clear policies to protect users from harm” while utilizing its AI tools. Editing features for modifying audio and speech are currently still under testing.
All content generated with Gemini Omni will incorporate Google’s imperceptible SynthID digital watermark for verification purposes. Users have expressed concerns over the “uncanny valley” effect seen in output quality from Veo 3.1, and it remains to be seen if Gemini Omni’s results will alleviate these issues.
Gemini Omni Flash is now accessible to Google AI Plus, Pro, and Ultra subscribers globally, with rollouts to users of YouTube Shorts and the YouTube Create App anticipated to begin this week.
