
Developers building creative tools
Ship Transition Studio-style apps: drop first and last frames, generate the in-between, or build a Draft Room that compares 360p variations side by side.
Google DeepMind's production-ready update: extend scenes up to 40 seconds, generate between first and last frames, iterate in 360p, and deliver 1080p or 4K. Built for directors, developers, and creative tools.
Generate a video to see the preview in this area
Omni brought real-world reasoning to generative video. 1.1 Flash makes it production-ready โ longer stories, keyframe control, faster drafts, and 4K finish.
Continue a clip from where it left off. Omni 1.1 analyzes up to 10 seconds of prior context โ not just the last second โ so characters, lighting, and narrative hold as you add 10-second increments up to 40 seconds total.
Drop in a start and end frame and Omni 1.1 generates continuous video between them. Ideal for camera orbits, zoom transitions, whip-pans, and seamless loops โ one shot, no jump cuts.
Generate lightweight 360p previews up to 60% faster and at about a third of the cost of 720p. Iterate on storyboards and variations cheaply, then commit to a higher resolution when the shot is right.
When a draft lands, generate polished 1080p or 4K output ready for professional production โ social, ads, and cinematic delivery without a separate upscaler.
Reference up to three seconds of video (plus images) so motion, characters, and style stay consistent. Combine text, images, and clips in one multimodal generation.
Use scene extension, first/last frames, and 360p drafts to iterate quickly, then deliver in 4K.
Go to the generator below to start a clip โ text prompt, reference images, first/last frames, or a short video reference.

Built for developers, directors, and teams who need controllable generative video โ longer scenes, real camera moves, and production-ready resolution.

Ship Transition Studio-style apps: drop first and last frames, generate the in-between, or build a Draft Room that compares 360p variations side by side.

Extend a take instead of regenerating from scratch. Keyframe interpolation gives you orbits, dolly-zooms, and continuous shots that read as directed, not sampled.

Walkthroughs, ads, and explainers that stay accurate under scrutiny โ then finish in 4K for social and campaign delivery.
Scene extension continues footage from where it left off. Omni 1.1 reads up to 10 seconds of prior context so characters, lighting, and narrative stay consistent as you add 10-second increments, up to 40 seconds total.
Gemini Omni 1.1 Flash is a multimodal video model โ text and video in/out, image in โ with native speech, music, and sound effects, plus C2PA Content Credentials.
Previous models often referenced only the last second. Omni 1.1 looks at up to 10 seconds of prior footage, so visual consistency and story beats survive each 10-second extension.
Generate continuous video between two keyframes. Complex camera orbits, zooms, and looping clips become a directed shot instead of a jump cut.
Prototype up to 60% faster at roughly one-third the cost of 720p, then output 720p, 1080p, or 4K when you are ready to ship.
16:9 and 9:16, up to 10 seconds per clip, 3 videos and 10 images per prompt, 360pโ4K, 131K input tokens. Model ID: gemini-omni-1.1-flash-preview.
Multiple AI video models, each built for different creative needs โ from Omni 1.1 Flash controls to cinematic storytelling.

Scene extension, first/last frames, 360p drafts, and 4K upscaling with multimodal text, image, and video input.

Google DeepMind's natively multimodal video model โ create and refine from text, image, audio, and video with iterative editing.

Short-form cinematic storytelling with native audio, multi-shot control, and up to 4K output.

Google's flagship video model for high-fidelity generation with strong prompt following and cinematic output.
The same controls Google shipped for developers โ scene extension, keyframes, and 4K โ ready to try in the generator.
Still have questions? We're here to help.