Gemini Omni Flash
2026-05-19 · Publisher : Google · Domain : Artificial Intelligence
Google introduces Gemini Omni at I/O, a model that creates from any input, starting with video: Gemini Omni Flash generates and edits videos through conversation, with reasoning about the real world.
What changes
- Combinable inputs: images, audio, video and text, to produce high-quality video.
- Video editing through dialogue, each instruction building on the previous ones, with consistency of characters and physics.
- Better intuitive physics (gravity, kinetic energy, fluids) and Gemini's world knowledge.
- Multiple references: image, text, video; for audio, only voice references at launch.
- Image and audio outputs planned for later.
### Specifications - Availability at launch: Gemini app, Google Flow and YouTube Shorts. - Developer access on June 30, 2026 (Gemini API, AI Studio, Gemini Enterprise Agent Platform) under the identifier gemini-omni-flash-preview, discontinued on September 30, 2026 in favor of gemini-omni-1.1-flash. - Context and maximum duration: not disclosed in the post. ### Pricing When it opened to developers (June 30), the post announces 0.10 USD per second of generated video, the same price as Veo 3.1 Fast. Pricing at the May 19 announcement: not disclosed. ### Benchmarks No score published in the post. ### Key takeaways Google is moving video generation toward a unified multimodal model, with conversational editing as the main contribution.
Links
Tags : Vidéo, Google, Gemini, Omni, Génératif