Gemini Omni announced at Google I/O 2026 — unified multimodal model accepts text + image + audio + video in one prompt

Google announced Gemini Omni at I/O 2026 (May 19) — a unified multimodal model that accepts text, image, audio, and video in a single prompt and reasons across all four modalities to produce a video output. The release positions Google as the lead in the all-in-one-model approach to multimodal generation.

The architectural bet is significant. Where Seedance, Veo, and Sora are increasingly orchestrated pipelines of specialized models (Vovoo on VO3 AI is the canonical example, routing between Veo 3.1, Sora 2, Kling 3.0, Seedance, Hailuo, Hunyuan, Nano Banana Pro per pipeline step), Gemini Omni is the counter-bet: one model that handles everything natively.

Both architectures will exist. The pipeline approach wins on best-in-class per step; the unified approach wins on consistency, latency, and cost. The Q3 enterprise question is which class of workload picks which. Long-form ad video probably stays pipeline-orchestrated; conversational multimodal probably moves to unified.

VO3 AI — Gemini Omni Google I/O 2026 → · WaveSpeed — Seedance vs Kling vs Sora vs Veo →