Set the first frame and the last frame, and let it generate the move between them.
Gemini Omni 1.1 Flash now on MITO
Gemini Omni 1.1 Flash by Google turns conversational video editing into shot planning. Set a start frame and an end frame and it generates the camera move between them. Feed it up to three video references to carry a character or a motion into a new scene. Draft in 360p at a third of the cost, then finish the take you keep in 4K, all inside MITO.
Real outputs and the prompts behind them. Remix any of them in one click.
One continuous cinematic shot, no jump cuts. Start on a close-up, low-angle view of a stylish drummer in a tailored beige suit, energetically playing a red drum kit, set within a grand ornate hall. The camera then whip-pans smoothly to the side in one fluid motion, revealing an older, distinguished saxophonist playing alongside him, and beside the saxophonist a ballet dancer in a flowing white outfit spinning gracefully. The whole scene is bathed in soft purple stage lighting, with dramatic shadows and atmospheric haze in the grand hall. The camera move should feel like one unbroken, fluid whip-pan reveal connecting all three performers, ending with all three visible together in a wide dynamic composition. No text or logos.
Try itLuxury, 8K ultra-realistic cinematic skincare commercial for "MITO ACTIVE SERUM," shot on a 35mm lens under high-end studio lighting against a dark pitch-black background. The featured product must match the reference bottle exactly: a rounded clear glass bottle holding warm amber-toned serum, a rose-gold pipette dropper cap, and a cream label reading "MITO" with smaller text "ACTIVE SERUM" beneath it in a clean serif/sans typographic style. Ignore the reference photo's grocery-tray, plastic-wrap, lemon, and price-tag styling entirely - only the bottle's shape, cap, glass tone, and label design carry over. Opens with a macro extreme close-up of the glass pipette dropper filled with viscous, faintly bubbly amber serum as a pristine drop slowly forms at its tip. Cut to slow-motion liquid physics as the drop falls and creates smooth, elegant circular ripples across a reflective black surface. Reveal a centered hero shot of the MITO Active Serum bottle standing on a glossy mirror surface. Cut to a macro close-up of a gorgeous East Asian model with dewy, radiant skin as the pipette applies serum to her cheek, followed by her fingers gently massaging the glowing serum into her skin; she smiles softly at the camera, holding the MITO bottle beside her face in soft-focus fashion lighting. Ends on a crisp, static high-fashion shot of the MITO bottle centered on the dark reflective surface. Shot at 60fps slow motion, smooth cinematic camera pans, dramatic rim lighting, soft key light, hyper-detailed skin textures, commercial advertisement quality. No on-screen text beyond what already appears on the bottle label.
Try itA premium 10-second cinematic product commercial for a wireless speaker, shot with deep blacks and moody, high-end studio lighting throughout. Open tight on the speaker's fabric mesh grille and brushed-metal top edge, with the LED base ring barely visible. Hold and slowly drift across this and one or two more deliberate cinematic close-ups (the glowing LED ring pulsing gently, the metal cap catching a sharp rim light against near-black shadow), emphasizing premium materials and craftsmanship. Then, in one continuous smooth cinematic move, the camera pulls back and reveals the full product. The ad ends on the final shot: the speaker alone, perfectly centered in frame, sitting on a glossy black reflective surface against a pure black background, calm and still. No on-screen text or logos.
Try itThe 1.1 controls are all about carrying intent between shots: a frame you pin, a reference you reuse, a take you re-run at higher resolution. That only pays off if your context survives between generations. On MITO it does. Your references, your scenes, and your previous takes sit on the same canvas, so refining is one more step in the project rather than a re-upload.
Set the composition you open on and the one you land on, and the model fills the move. On MITO the reference frames, the clip, and the prompt history are already on the canvas, so you iterate from where you are instead of re-assembling the shot every time.
Character boards, style frames, and the short clips you use as motion references stay loaded. Point at them by @tag, switch models mid-project, and keep the same character across every shot.
Block the shot in 360p, iterate until the timing is right, then re-run the take you like at 1080p or 4K. Same canvas, same prompt, no exporting between tools.
One model, many workflows. Here is how different teams put it to work.
Pin the first and last frame when a shot has to land on an exact composition, then match a new shot to footage you already have by attaching it as a video reference.
Draft twenty versions of a hook in 360p for the price of six at 720p, pick the one that works, and re-run only that take at full resolution.
Attach the client's existing footage as a reference so a new shot matches the campaign already in market, and pin the opening frame when it has to sit in a cut you have already locked.
Generate, then refine. It is never one and done.
Video: Google
Write the first ten seconds as you would describe it to a DP: subject, camera position, light, and the movement you want. Attach a reference clip if a character or a motion has to match something you already have.
Video: Google
Give it the composition you open on and the one you land on, and it generates the continuous move between them. Attach up to three video references if a character or a motion has to match footage you already have.
Video: Google
Change anything by describing it, or pin a start and end frame to force a specific composition. Iterate in 360p while the idea is still moving, then re-run the take you keep at 1080p or 4K.
More power, more control, from prompt to video generation
Video: Google
Give it the composition you open on and the composition you land on. Gemini Omni 1.1 Flash generates the continuous motion in between, so a push-in, an orbit, or a rack across a room arrives framed the way you planned. Useful when a shot has to cut cleanly into an edit that already exists.
Start creatingVideo: Google
Attach up to three video references and the model pulls the character, the style, or the movement pattern into a new scene. It is the fastest way to match a shot to footage you already have, whether that is your own earlier take or the campaign the client shot last year.
Start creatingVideo: Google
Draft mode renders at 360p up to 60% faster and for a third of the cost of 720p. Use it to test timing, blocking, and camera direction while the idea is still moving. When a take works, re-run it at 720p, 1080p, or 4K. You spend the expensive resolution on the version you are keeping.
Start creatingVideo: Google
The conversational editing that made Gemini Omni useful is all still here. Describe the change you want in plain language and keep going, turn after turn, with the scene context and the physics carrying across every refinement. No keyframes, no rotoscoping, no starting the generation over.
Start creating
Gemini Omni 1.1 Flash is Google's multimodal video model, generally available since 27 August 2026. It keeps the conversational editing of Gemini Omni and adds keyframe control, video references, a cheap draft resolution, and 4K output.
| Specification | Value |
|---|---|
| Max Clip Length | 10 seconds per generation |
| Max Resolution | Up to 4K (360p, 720p, 1080p, 4K) |
| Draft Mode | 360p, up to 60% faster and 1/3 the cost of 720p |
| Keyframe Control | Start and end frame, continuous move generated between |
| Video References | Up to 3 per prompt |
| Input Types | Text, Image, Video, Audio |
| Audio | Native synced audio, same generation pass |
| Editing | Multi-turn conversational, plain-language instructions |
| Watermark | SynthID |
| Aspect Ratios | 16:9 and 9:16 |
| Credits per Generation | 40 to 320 credits per second, by resolution |
| Provider |
Every model lives on MITO. Pick by the job, not the name.
Three things you can use on MITO today. Keyframe control: you set a start frame and an end frame and it generates the camera move between them. Video references: attach up to three clips to carry a character, style, or motion into a new scene. And a 360p draft mode that runs up to 60% faster at a third of the cost of 720p, with output up to 4K when you are ready to finish. The conversational, any-input editing from Gemini Omni is all still there. Google's 1.1 release also adds scene extension, which is not wired up on MITO yet.
Ten seconds per generation on MITO. If you need a single shot that runs longer, Seedance 2.5 generates up to 30 seconds in one pass and is also on MITO. For a sequence, generate the shots you need and cut them together on the canvas, where your references and previous takes stay loaded between generations.
Not on MITO yet. Google's 1.1 release adds scene extension in 10-second steps, but it is not wired up on MITO at the moment, so generations run to 10 seconds. Everything else in the release is available today: keyframe control, video references, draft mode, and 4K output. If you need a longer single take right now, Seedance 2.5 on MITO generates up to 30 seconds in one pass.
Yes. Give Gemini Omni 1.1 Flash the frame you want to open on and the frame you want to land on, and it generates the continuous motion between the two. That covers push-ins, orbits, and transitions where the final composition matters, for example when the shot has to cut into a sequence that already exists.
Draft mode generates at 360p, up to 60% faster and for about a third of the cost of a 720p generation. Use it while the idea is still changing: testing timing, blocking, camera direction, or which of six prompt variations reads best. Once a take works, re-run the same prompt at 720p, 1080p, or 4K. You only pay full price for the version you keep.
On MITO the model sits on a canvas where your references, previous takes, and scene context are already loaded and travel with every generation. Pinning a frame or matching a reference is one more step in the project. Going direct through an API or a standalone tool, you rebuild that context each time. MITO also puts storyboards, scene breakdown, editing, and every other model next to the generation, so a handful of shots can become a finished cut without changing tools.
Gemini Omni introduced conversational, any-input video creation and editing: start from text, image, video, or audio and refine by describing the change. Gemini Omni 1.1 Flash keeps all of that and adds the controls that matter once a shot has to fit into a real edit: start and end frame control, video references, a cheap draft resolution, and 4K output. Both are on MITO, so you can use the original for open-ended exploration and 1.1 Flash when the shot has constraints.
They solve different problems. Seedance 2.5 generates up to 30 seconds in a single pass with up to 50 references, so it is the choice when a shot has to run long. Gemini Omni 1.1 Flash works in 10-second takes but gives you precise control over them: pin the opening and closing frame, carry a character in from a reference clip, and iterate cheaply in draft before committing to 4K. Kling 3.0 leads on natural motion and physics for shorter clips. All three are on MITO, so you pick by what the shot needs.
Yes. Videos you generate with Gemini Omni 1.1 Flash through MITO can be used commercially, including advertising, social campaigns, short films, and client deliverables. Output carries Google's SynthID watermark, which identifies the content as AI-generated and does not restrict commercial use. Check MITO's terms of service for the full licensing detail.
Generations cost 40 to 320 credits per second depending on the resolution you pick, so a 10-second clip ranges from 400 to 3,200 credits. Drafting at 360p costs roughly a third of a 720p generation, so the cheapest way to work is to iterate in draft and spend full resolution only on final takes. Start with free credits to run your first generation. No card needed.
Available on all MITO paid plans. Sign up today.
Start creating