Google launches Gemini Omni to simulate reality with multimodal AI

Google has launched Gemini Omni, a new family of multimodal models designed to create content from any input. The first model in the family, Gemini Omni Flash, is rolling out today to the Gemini app, YouTube Shorts, and AI creative studio Flow. The launch represents a strategic shift for the company, which first announced Gemini three years ago. Google CEO Sundar Pichai explained that when Gemini was first announced, the goal was to build a natively multimodal model trained on text, code, audio, images, and video to give it a deeper understanding of the world. Pichai noted that with world models, AI is moving from predicting text to simulating reality, and Gemini Omni represents the next step in that direction.

The Gemini Omni Flash model is capable of rendering 10 seconds of video. Nicole Brichtova, Google DeepMind director of product management, described this limitation as a strategic choice for consumer accessibility rather than a technical constraint, noting that the company wanted to get the tool into more hands. Google is positioning Omni Flash primarily as a consumer tool. Gabe Barth-Maron, a research engineer at DeepMind, described the outputs by saying, “They’re like personalized memes.” Brichtova explained that the release represents the next step in combining the intelligence of Gemini with the rendering capabilities of the company’s media models, distinguishing Omni from Google’s existing video model, Veo.

Google plans to make Omni available via an Application Programming Interface (API) in the coming weeks. The technology allows users to generate videos using their own digital avatars. However, to prevent deepfakes, users will have to go through a dedicated product onboarding. This process requires users to record themselves and speak a series of numbers before the avatar is stored. Additionally, all videos created with Omni will include Google’s SynthID digital watermark to verify if they were generated via Gemini products. The launch comes as competitors like Luma AI build similar agentic tools. During a media briefing on Monday, Google executives detailed the strategic vision behind the launch, emphasizing the transition from simple text prediction to simulating reality.

Why it matters

Google’s release of Gemini Omni represents a concrete step toward the company’s goal of building a multimodal large language model that can reason across text, image, audio, and video to simulate reality.