Overview
- Google DeepMind announced Gemini Omni 1.1 Flash Thursday, adding 1080p and 4K upscaling to the Omni Flash video model that began life as a 720p clip generator.
- The release introduces scene extension that can analyze up to 10 seconds of prior context and extend clips in 10-second steps up to a 40-second total length.
- Developers can use conversational, multi-turn editing with multimodal inputs via the Gemini API and Google AI Studio to refine video through iterative prompts.
- All Omni outputs include Google’s SynthID watermarking to label AI-generated media, and the features are rolling out into Google Flow, the Gemini app, and enterprise deployment paths for subscribers.
- Key gaps remain: public API docs, exact pricing for 1080p and 4K tiers, and some outlets note discrepancies between Google’s materials and secondary reports so teams should verify availability before building production workflows.