Google’s Gemini Omni turns images, audio, and text into video — and that’s just the start

Google has introduced Gemini Omni, a new suite of multimodal AI models unveiled during its Google I/O developer conference, marking a significant advancement in the company’s long-term vision for unified AI systems. The initiative builds on Google’s original Gemini project, launched three years ago with the aim of creating a single neural network capable of understanding and generating text, images, audio, and video.

According to Google CEO Sundar Pichai, Gemini Omni is designed to “create anything from any input,” allowing users to combine multiple media formats — including text, images, audio, and video — into a single workflow. Rather than simply merging these inputs, the AI model processes and reasons across them to produce coherent, high-quality outputs with contextual understanding across science, history, culture, and physics.

The first rollout of Omni will focus primarily on video generation. Users can create AI-generated videos using multimodal prompts, while also editing photos using natural language instructions instead of relying on advanced editing tools. The functionality builds upon Google’s earlier AI media initiatives, including its Veo video-generation platform and Nano Banana image-editing technology.

Nicole Brichtova described Gemini Omni as a major evolution beyond existing video models, explaining that it combines Gemini’s intelligence capabilities with advanced media rendering systems developed by Google DeepMind.

During a media briefing, DeepMind chief technologist Koray Kavukcuoglu demonstrated Omni’s capabilities using prompts such as a claymation-style explanation of protein folding, which the model transformed into a fully narrated stop-motion educational video within seconds.

Google’s long-term strategy for Omni extends beyond video generation. Future capabilities may include generating images from sound inputs or creating audio directly from videos, moving AI closer toward “world models” capable of simulating real-world understanding rather than only predicting text responses.

To address growing concerns around deepfakes and synthetic media, Google announced several safety measures. Users creating personalized digital avatars will need to complete identity verification during onboarding, including video recording steps. In addition, all AI-generated videos from Omni will carry Google’s SynthID watermark technology to help identify AI-created content.

The first version, Gemini Omni Flash, is launching across the Gemini app, YouTube Shorts, and Google’s Flow creative studio. Initially, the model will support videos up to 10 seconds long, though Google confirmed longer-duration video generation is planned in future updates.

Google appears to be positioning Omni Flash primarily toward mainstream consumers and creators. Early use cases include generating personalized videos, editing travel clips, and creating avatar-driven content for social media. Researchers at DeepMind described these experiences as highly personalized and consumer-friendly, aiming to make AI video creation more accessible to general users.

At the same time, the company sees strong commercial and enterprise potential for Omni. Google plans to release API access in the coming weeks, enabling developers, advertisers, filmmakers, and content creators to integrate multimodal AI workflows into professional production pipelines.

The launch also highlights intensifying competition in the generative AI space, where companies including OpenAI and startup Luma AI are developing similar AI-powered media generation systems.

Google also revealed that a more advanced version, Gemini Omni Pro, is currently under development. While no official release date has been announced, the company stated the Pro version will target more demanding enterprise and professional use cases with enhanced performance across all Omni functionalities.

Voice Of Osiz

Osiz Technologies believes Google’s Gemini Omni launch marks a major leap in the evolution of multimodal AI and intelligent content creation. The ability to generate videos, images, audio, and text through a unified AI system highlights how rapidly AI is transforming digital experiences across industries. As enterprises increasingly seek automated creative workflows, multimodal AI models are expected to reshape media production, marketing, advertising, and personalized user engagement. The integration of advanced reasoning with AI rendering capabilities also opens new opportunities for businesses to build smarter and more interactive applications. Features like AI avatar generation, real-time editing, and multimodal understanding demonstrate the growing potential of next-generation generative AI platforms. At Osiz, we see this advancement accelerating demand for AI development, automation solutions, and enterprise-grade creative AI systems. Businesses adopting these innovations early will gain a strong competitive edge in the evolving digital economy.

Source: Techcrunch.com