Google Unveils Gemini Omni Multimodal Video Model With Conversational Editing

Google has released Gemini Omni 1.1 Flash, expanding its generative-video lineup as competition shifts from producing isolated clips to supporting longer, more controllable sequences. The model targets a persistent challenge in AI video: keeping characters and scenes consistent across separately generated segments. First- and last-frame controls, alongside tiered resolution pricing, are designed to give creators more influence over shot composition, output quality and production costs.

The upgrade adds a scene-extension feature that can build continuous, segmented videos lasting as long as 40 seconds. It also increases the preceding context used for each continuation to 10 seconds, aimed at improving character consistency between clips. Gemini Omni 1.1 Flash supports specified opening and closing frames, four resolution-based pricing tiers and 4K upscaling. The supplied reports did not specify a launch date or disclose prices for the four tiers.