Gemini Omni is Google’s first any-to-any multimodal “world model.” You give it any combination of text, image, audio, and video, and it generates video grounded in real-world knowledge. The first model, Gemini Omni Flash, is shipping now. ✍️
Text → video. Describe a scene; get a 10-second clip with synced audio. 🖼️
Image → video. Upload a photo — a product, a person, a sketch — and bring it to life.🎞️
Video → video. Restyle, edit, or add effects to footage you already have.🎵
Audio → video. Upload a song or voiceover and get video that matches its pacing, emotion, and beats — genuinely unique to Omni. 💬
Conversational editing. Refine a clip by chatting: “make the lights dimmer,” “remove the violin,” “change the angle.” Every instruction builds on the last, and the scene remembers what came before. 🌍
World-model physics. It has an intuitive grasp of gravity, kinetic energy, and fluid dynamics — so a marble rolling down a track bounces and sounds right.
Think of it less as a video generator and more as a creative collaborator that already understands how the world looks, moves, and sounds.
