Gemini Omni, Google’s new any-input AI model, is now live in Picsart. It arrives across four places creators already work: the AI Playground, the AI Video Generator, the AI Video Editor, and Picsart Flow. The model was built to turn images, audio, video, and text into finished video, and now that capability sits inside the same Picsart workspace where ideas already become designs, edits, and posts.
This is the first time a single model lets a creator move from a rough reference to a polished clip without leaving the canvas. Below is what Gemini Omni does, where to find it in Picsart, and how to put it to work today.
Meet Gemini Omni
Gemini Omni is Google’s multimodal model designed to “create anything from any input, starting with video,” according to Google’s announcement. What sets it apart from earlier Gemini releases is that it combines reasoning with creation in one model, so it draws on Gemini’s knowledge of history, science, and cultural context while it generates.
The result shows up in the details. Google notes the model has an improved intuitive understanding of forces like gravity, kinetic energy, and fluid dynamics, which is exactly what makes generated motion read as believable rather than uncanny. The first release in the family is Gemini Omni Flash, tuned for speed.
What Gemini Omni can do
Gemini Omni takes images, audio, video, and text as inputs and produces video, with image and audio outputs planned over time. That flexibility is the headline: a single still, a voice clip, a reference video, or a written line can all become the starting point for a finished scene.
Editing is conversational. Google describes natural language video editing where “every instruction builds on the last,” so a creator can refine a clip across multiple turns and keep character consistency, physics accuracy, and scene continuity intact along the way. Instead of restarting from scratch with each change, the model treats the project as one continuous conversation.
Gemini Omni in Picsart AI Playground
The Picsart AI Playground is where Gemini Omni is easiest to test. The Playground is built for experimenting with models side by side, so it is the natural place to feed in a reference image or a text prompt and see what Omni returns before committing it to a larger project.
This matters for anyone weighing models against each other. With Omni now in the Playground, a creator can compare its output to the Veo and Nano Banana models already there and pick the right tool for each brief, all on one Picsart subscription.
Gemini Omni in the Picsart AI Video Generator
For finished clips, Gemini Omni runs inside the Picsart AI Video Generator. This is the surface for turning a prompt or a starting image into video, and Omni’s any-input approach fits it directly: describe the scene, drop in a reference, and let the model handle the motion.
Because the model brings a stronger grasp of real-world physics, the generator becomes useful for the moments that usually break AI video, like falling objects, flowing liquid, and natural movement. Those are the shots that sell a clip as real, and they are the ones Google built Omni to handle.
Conversational editing in the Picsart AI Video Editor
The Picsart AI Video Editor is where Omni’s multi-turn editing comes to life. Rather than dragging keyframes, a creator can describe the change in plain language and refine it step by step, with each instruction building on the last.
Continuity is the payoff here. Because the model holds character consistency and scene continuity across edits, a series of small adjustments stays coherent instead of drifting into a different look with every pass. That makes the editor practical for real revision cycles, not just one-shot generations.
Gemini Omni in Picsart Flow
Inside Picsart Flow, Gemini Omni becomes a building block in a larger creative pipeline. Flow is built to chain steps together, so Omni’s generation and editing can sit alongside other Picsart tools in a repeatable workflow rather than a single isolated render.
For teams producing video at volume, that is the difference between a one-off experiment and a process. A reference asset can move through generation, editing, and finishing in one connected flow.
Built-in transparency
Every video Gemini Omni creates carries an imperceptible SynthID digital watermark, Google’s system for identifying AI-generated content. That verification travels with the output, so creators and audiences keep a reliable signal of how a clip was made.
Get started
Gemini Omni is live in Picsart now. Open the AI Playground to test it, jump into the AI Video Generator to build a clip, refine it in the AI Video Editor, or wire it into a repeatable pipeline with Picsart Flow. Pick a starting input and turn it into video today.
