Google unveils Gemini Omni avatars

Google used its I/O 2026 conference on May 19 to introduce Gemini Omni, a new family of generative AI models built to create video from combinations of text, images, audio and video. In a company blog post, Google said the first release, Gemini Omni Flash, is rolling out to the Gemini app, Google Flow and YouTube Shorts. CNN Business reported on May 23 that the new capability can generate realistic avatars of users, adding a more consumer-facing use case to Google’s broader push to put Gemini across search, apps and devices. The announcement came alongside a wider set of Gemini updates at I/O, including Gemini 3.5 Flash and Google Antigravity, which the company described as tools aimed at more agentic AI tasks. Separately, TechRadar reported on May 24 that Google also pushed a fix for first-generation Chromecast dongles and a Gemini-related upgrade to the 4K Chromecast with Google TV. Together, the announcements showed Google tying its latest AI models to both creation tools and consumer hardware. ### What exactly did Google say Gemini Omni does? Google said on May 19 that Gemini Omni can “create anything from any input — starting with video.” In its product post, the company said users can combine images, audio, video and text as inputs and generate videos that draw on Gemini’s real-world knowledge. The same post said Gemini Omni also supports conversational editing, letting users change videos through natural-language prompts rather than traditional editing controls. Google said image and audio outputs would be added later, indicating that the first phase is centered on video. CNN Business reported on May 23 that one feature shown by Google allows people to generate realistic avatars of themselves. That report framed the avatar capability as part of a broader shift in how AI systems are changing search and online content creation. ### Where is Google putting the new model first? Google said Gemini Omni Flash is rolling out first in three places: the Gemini app, Google Flow and YouTube Shorts. That distribution matters because it places the model in both a general-purpose AI assistant and video-focused products. Google’s I/O roundup also listed Gemini Omni among the company’s headline announcements from the conference. In a separate keynote post, Chief Executive Sundar Pichai said Google was bringing Gemini-powered experiences across products including Search. ### How does this connect to Gemini 3.5 Flash, Spark and Antigravity? Google said Gemini 3.5 Flash is the first model in a new family designed to combine “frontier intelligence with action.” In its model post, the company said the system is built for long-horizon agentic tasks and that Gemini Spark, described as a personal AI agent, uses 3.5 Flash. Google’s developer highlights post said Antigravity is its “agent-first development platform.” PPC Land described Gemini 3.5 Flash, Spark and Antigravity as part of an I/O push toward agentic AI, with systems intended to move beyond answering prompts and toward carrying out multi-step work. ### Why are Chromecast devices part of the same week’s story? TechRadar reported on May 24 that Google rolled out a fix for first-generation Chromecast dongles after users said the devices had stopped working. The same report said the 4K Chromecast with Google TV received a major Gemini upgrade. TechRadar said the updated Chromecast interface includes a Gemini screen for image creation and voice prompts similar to Gemini on the web and mobile. That report also said some apps no longer support the oldest Chromecast hardware. ### What comes next from here? Google said future Gemini Omni releases will add output modalities including image and audio, beyond the current video-first launch. The company also said Gemini Omni Flash is already beginning to roll out across the Gemini app, Google Flow and YouTube Shorts, while Gemini 3.5 Flash and Antigravity are being positioned as the engine for newer agent features announced at I/O.