Any input, one model
Text, a reference image, an audio track, an existing clip: all of it goes into one prompt and the model reasons across the lot. You stop assembling a pipeline of specialised tools to get one result.
Gemini Omni is Google's unified multimodal model, announced at Google I/O in May 2026. It takes text, images, audio, and video in a single prompt and reasons across all of them to produce video with synchronized sound.
The distinction from Veo is architectural. Instead of chaining a video model to an image model to an audio model, Omni handles the whole thing inside one Gemini-family system, which is why audio matches what is happening on screen rather than being layered on afterwards.
Short social clips, product moments, concept scenes, and reference-led video
Video, social, and brand teams working from mixed inputs rather than a text prompt alone
Mixed inputs, native synchronized audio, and one model doing the whole job.
Text, a reference image, an audio track, an existing clip: all of it goes into one prompt and the model reasons across the lot. You stop assembling a pipeline of specialised tools to get one result.
Audio is generated in the same pass as the picture, so the model works out what should be audible from what is visually happening. Ambience and effects fit the shot instead of being laid over it afterwards.
Rather than chaining a video model, an image model, and an audio model together, Omni does the reasoning and the generation inside one Gemini-family system. Fewer joins means fewer places for the result to come apart.
Refinement happens in natural language rather than by re-rolling the whole generation with a reworded prompt. Adjusting a clip becomes a conversation about it.
Google has framed the family as creating anything from any input, starting with video, with image and audio output planned for later models. In Phygital+ it sits on the same canvas as 30+ other tools, so it slots into workflows you already have.
Gemini Omni fits short-form work where the inputs are mixed and the output needs sound: social clips, product moments, and concept scenes.
Give it the inputs you have, describe the scene, and generate video with sound.
Give the model the strongest inputs you have. A reference image anchors the look, an audio track sets the rhythm or the voice, and the text prompt covers the scene, the motion, the camera behaviour, and the pacing.
Describe what should move and what should stay stable, and name the lighting, mood, and pacing. Clips are short, so a single clear beat works better than a sequence compressed into ten seconds.
Review the native audio alongside the picture, since both come from the same pass. Chain the clip into a video upscaler for delivery resolution, or into an audio node to layer additional sound design, then export.
One subscription across 30+ AI models, with no per-tool credit balances or separate signups. Credit cost per generation is shown live in the node before you run it.
500 weekly credits to test the model and see the output quality. No credit card required.
45,000 monthly credits with cheaper per-credit pricing, commercial rights, and video download.
90,000 monthly credits, up to 10 seats, shared workspace, centralized billing, and priority support.
210,000 monthly credits, API and integration support, unlimited seats, and dedicated management.
Full specs for Gemini Omni in Phygital+.
How Gemini Omni compares with other AI video models in the Phygital+ catalog.
Chain image generation, video, upscaling, and audio into one repeatable workflow.
Generate a reference frame with an image model, then hand it to Gemini Omni as visual input for the clip.
Take the finished clip into a video upscaler to reach delivery resolution for ads or broadcast.
Generate narration or a music bed in an audio node and use it as input, or layer it over the native audio.
Write the scene description and shot list with a text node before sending it into the prompt.
Everything you need to know about Gemini Omni in Phygital+.
Gemini Omni is Google DeepMind's unified multimodal model, announced at Google I/O in May 2026. It accepts text, images, audio, and video in a single prompt and reasons across all of them to produce one output, starting with video. The first model in the family is Gemini Omni Flash.
Veo is a dedicated video generation engine. Omni is a single Gemini-family model that handles the reasoning and the generation together, rather than chaining a video model, an image model, and an audio model into a pipeline. In the Gemini app, Omni Flash replaced the earlier Veo-powered flow.
Every clip comes with synchronized audio generated in the same pass rather than added afterwards. The model reasons about what sound belongs in the scene based on what is visually happening, so ambience and effects match the shot instead of being layered over it.
Text, images, audio, and video, in any combination, within one prompt. That is the point of the architecture: a reference photo, a voiceover track, and a written brief can all inform the same generation without being fed through separate tools first.
Clips are short, on the order of ten seconds. For longer sequences, generate several clips and assemble them, or use a model built for scene extension.
Google has described the family's ambition as creating anything from any input, starting with video. Omni Flash shipped with video output first, with image and audio output planned for later models in the family.
Give it the strongest input you have rather than the longest description. If you have a reference frame, connect it. If you have a voice track, use it. Then describe the motion, the camera behaviour, and the pacing, and let the model reason across the rest.
Usage is credit-based and included in every plan, from Free up to Enterprise. The credit cost per generation is shown in the node before you run it, so there is no separate Google AI subscription to manage.
Outputs created on any paid plan come with commercial-use rights. Check the model card for provider-specific limits before publishing.
Browse the full catalog of 30+ AI models available in Phygital+.
Join 100+ teams using Phygital+ – every model in one workspace
Try Phygital+ free