Gemini Omni Flash: Google's New Unified Multimodal Generation Model

CNBC·medium signal

Also announced at I/O 2026, Gemini Omni is a new model series combining Gemini's reasoning with native creation capabilities — accepting image, audio, video, and text input while outputting video grounded in real-world knowledge with easy editing. This marks Google's entry into unified multimodal generation that competes directly with OpenAI's Sora-class capabilities, with Gemini Omni Flash available now.