Google has released Gemini Omni 1.1 Flash, a production-oriented update to its generative video model with controls for extending scenes, specifying first and last frames, drafting at lower resolution, upscaling to 4K, and using video references in prompts.
The model is rolling out through Google AI Studio and the Gemini Enterprise Agent Platform. Google positions it as a developer release for creative tools and media workflows, not just a consumer video generator.
That positioning matters. The generative video market is shifting from impressive single clips toward systems that let teams direct, iterate, compare, and reuse assets without restarting from scratch. Omni 1.1 Flash is Google’s answer to that workflow problem.
The new controls are about continuity
Google says scene extension lets a user continue an existing video from where it left off. Omni 1.1 can analyze up to 10 seconds of prior context, compared with previous models that referenced only the final second, and extend videos in 10-second increments up to a cumulative length of 40 seconds.
The release also adds first-and-last-frame interpolation. Developers can provide a starting frame and an ending frame, and the model generates the transition between them. That is useful for camera moves, loops, and continuity-sensitive edits where a normal text prompt is too loose.
Google also says Omni 1.1 supports up to three seconds of video reference in multimodal input. The goal is to preserve visual context, character consistency, or motion references while generating a new clip.
Draft mode changes the cost of iteration
The most operator-useful detail may be draft generation. Google says developers can generate 360p previews up to 60% faster and at one third of the cost compared with Omni 1.1’s standard 720p resolution.
That is a practical workflow change. Video generation often takes several attempts before a clip is usable. A cheap draft loop lets teams test camera direction, motion, timing, and composition before spending on a polished render.
Google also says Omni 1.1 can generate 1080p and 4K outputs for final production. The announcement includes examples from Adobe Firefly, Figma Weave, GMI Cloud, and Runway, framing the model as something other creative platforms can embed rather than only a Google-owned app feature.
The model ID moves video into the API surface
Google’s example code uses gemini-omni-1.1-flash through the Interactions API. That matters because developer adoption depends on whether model controls can be wired into products, review flows, asset libraries, and billing systems.
The previous wave of AI video tools was often prompt-first. This release points to a more structured layer: scene continuation, keyframe control, low-resolution drafts, final upscales, and references as explicit workflow steps.
