Google Launches Gemini Omni - Nano Banana but for Video, Live in the Gemini App Today

Google has launched Gemini Omni, a new multimodal AI model that generates video content from any input type including text, images, and video. The model is available immediately within the Gemini App, Flow, and YouTube, marking Google's latest push into the competitive video generation space. Logan Kilpatrick described Omni as "Nano Banana but for video", referencing Google's viral image generation hit from last year. The comparison suggests Google aims to replicate Nano Banana's conversational and instant user experience for video creation. The model promises the same accessible, engaging interface that made its predecessor popular. The strategic significance lies in Google's unified approach. Rather than developing separate specialist models for different tasks, Omni positions itself as a single model capable of handling multiple input modalities whilst producing video output. Additional output formats are promised in future updates. This puts Gemini Omni in direct competition with established video generation platforms like Runway and Pika, as well as OpenAI's Sora. Google's bet centres on the advantages of a unified model over multiple specialists. The company believes that one comprehensive model can outperform dedicated tools by offering seamless integration across different content types. API access is planned but lacks a firm timeline, limiting current usage to the consumer-facing Gemini App interface.