Glonce

Google ships Gemini Omni 1.1 Flash with 40-second scene extension and 4K upscaling

modelsThe Glonce editorsHacker News🔥77Glonce rating77Impact68Novelty62Relevance98Surprise42Players88Credibility92
Google ships Gemini Omni 1.1 Flash with 40-second scene extension and 4K upscaling

Google has introduced Gemini Omni 1.1 Flash, a new suite of creative controls and generative video capabilities built on its Gemini Omni model. Google says the update, rolled out through the Gemini API in Google AI Studio, makes Omni 1.1 production ready for professional use, and is meant to make generative video more controllable, faster to iterate on, and polished for deployment in generative video workflows, creative tools, and media editing software.

The most significant change is scene extension. Omni 1.1 can now analyze up to 10 seconds of prior context when continuing an existing video, compared with earlier models that referenced only the video's final second. Google says the wider context window produces improved visual consistency and narrative adherence, letting developers build longer stories or branch a scene in new directions. Videos extend in 10-second increments, up to a cumulative total of 40 seconds.

A second new control lets developers specify a video's first and last frame; Omni 1.1 then generates continuous footage between the two keyframes. Google positions this for smooth transitions and camera movement, such as complex camera orbits, zoom transitions, or seamless looping clips.

For iteration, a new 360p draft mode renders lightweight previews. Google says it is up to 60% faster than Omni 1.1's standard 720p resolution, a figure the company attributes in a footnote to system throughput of 360p versus 720p, and that it costs a third as much as 720p generation. Google recommends the mode for rapid prototyping, storyboard iteration, and quick rendering on developer platforms.

At the other end of the pipeline, Omni 1.1 can upscale outputs to 1080p or 4K for finished, professional-grade production. The model also accepts video references as multimodal input: developers can supply up to three seconds of reference footage to keep visual context and character appearance consistent across a generated scene. Google's own example combines three separate dance reference videos with three character images, a dog, an octopus and a bear, to have the resulting characters perform the matching dances together in one continuous shot with no scene cuts.

Omni 1.1 is available now through the Gemini API in Google AI Studio, with documentation, a cookbook and prompting guides covering scene extension, video references and upscaling. Enterprises can build with it through the Gemini Enterprise Agent Platform API; Google says customers are already using that route for real-world production, though it names none of them. Separately, Omni 1.1 is rolling out today to all Google AI Plus, Pro and Ultra subscribers globally in Google Flow, and scene extension specifically is rolling out to the same subscriber tiers in the Gemini app.

Key facts

  • Scene extension now analyzes up to 10 seconds of prior context, versus earlier models that used only the final second, and videos can extend in 10-second increments to a cumulative total of 40 seconds.
  • A new keyframing mode generates continuous video between a specified first and last frame, aimed at camera orbits, zoom transitions and seamless loops.
  • A 360p draft mode renders up to 60% faster, based on system throughput versus 720p, and costs a third as much as Omni 1.1's standard 720p resolution.
  • Omni 1.1 can upscale outputs to 1080p or 4K, and accepts up to three seconds of reference video to keep characters and visual context consistent across a scene.
  • The update is live via the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform API, and Omni 1.1 (with scene extension specifically in the Gemini app) is rolling out today to all Google AI Plus, Pro and Ultra subscribers globally in Google Flow.

Why it matters

Gemini Omni 1.1 Flash is a jump in how much control developers get over generative video, not just how good one clip looks. The scene-extension window grows tenfold, from referencing only a video's final second to analyzing a full 10 seconds of prior context, and clips can now be chained in 10-second increments to a cumulative 40 seconds, closer to an actual scene than a single demo loop. Paired with first and last frame keyframing for controlled camera moves and transitions, and with 4K upscaling for finished output, Google frames Omni 1.1 as production ready for professional use rather than an experimental demo, which is the gap that has separated most generative-video showcases from real production pipelines.

Who it affects

The update targets developers building generative video workflows, creative tools, or media editing software through the Gemini API in Google AI Studio. Enterprises get a separate route through the Gemini Enterprise Agent Platform API, and Google says customers are already using it in production, though none are named. Individual users are affected too: Omni 1.1 is rolling out today to every Google AI Plus, Pro and Ultra subscriber globally inside Google Flow, and the scene-extension feature specifically is rolling out to the same subscriber tiers inside the Gemini app.

How to use it

Developers reach Omni 1.1 through the Gemini API, referencing the model as gemini-omni-1.1-flash and, for scene extension, passing a previous_interaction_id to continue an earlier generation; a response_format field sets output resolution, including the new 360p draft option. Google AI Studio, official documentation, a cookbook and prompting guides cover integrating scene extension, video references and upscaling. Enterprises use the Gemini Enterprise Agent Platform API instead, and subscribers to Google AI Plus, Pro or Ultra get Omni 1.1 in Google Flow, plus scene extension in the Gemini app, without writing any code. The source references a pricing table for Omni 1.1 as an image but states no pricing figures in the article text itself; the only cost detail given is that 360p draft previews cost a third as much as standard 720p generation.

How solid is it

This is Google's own launch announcement for its own model, written in the company's collective voice; no individual author, engineer or executive is named or quoted anywhere in it. The customers Google says are already running Omni Flash in production are not identified, and no competing model or product is named for comparison. The one performance figure given, that 360p previews render up to 60% faster than 720p, carries a footnote attributing it to system throughput of 360p versus 720p resolution. No technical explanation is given for how the model analyzes 10 seconds of prior context.

Risks and caveats

No pricing figures appear in the article text: the only cost information is that 360p draft generation costs a third as much as standard 720p, with no dollar figures for any resolution or tier. The 40-second maximum is a cumulative ceiling reached by chaining 10-second extensions, not a single continuous generation. No release date is given beyond the article's repeated use of the word today. The announcement names no specific customer and makes no comparison with any competing generative-video model or tool.