Gemini's new video model, 'Omni,' has been discovered by users ahead of its official launch, with reports surfacing on Reddit about its capabilities. Users noted that the model, labeled "Powered by Omni," appeared in the Gemini app, offering advanced video generation features. One user praised Omni for its prompt adherence and seamless multi-camera transitions, highlighting its superior audio and ambient sound quality compared to the existing Veo series. Despite the praise, users reported strict rate limits, with Pro subscribers exhausting 80% of their quota after generating just two videos. Additionally, safeguards still block celebrity likenesses, as tests with Will Smith's likeness failed. The discovery suggests Google may be integrating text, image, and video generation into a single architecture, aligning with DeepMind CEO Hassabis's previous statements about merging Gemini and Veo. An official announcement is anticipated at the Google I/O conference on May 19.