May 20, 2026 today Google I/O CONFERENCE OF DEVELOPERS, GOOGLE OFFICIALLY RELEASED Gemini Omni Model.
Demis Hassanis, Nobel Laureate and Head of Google DeepMind, stated that Gemini Omni was the most capable version of the Gemini model family to date。"Omni" means "All Power", the model displays unprecedented fluidity and depth when dealing with a variety of modular information, such as text, images, videos and audio。
Gemini Omni claims to be able"Generate any output from any input”Speech editing is supported, and the user can change the role, background, etc. in the video in one sentence。
1 From the General Assembly, Google also announced the launch of the Gemini Omni family head model Gemini Omni Flash, which is available on Gemini App, Google Flow, YouTube Shorts, and will push API services in the future。