Gemini Omni Flash
Gemini Omni Flash is a video model family by Google. Vini uses the current Omni Flash family for video generation, references and editing. A video reference can influence appearance and filming, but a style input is not an exact camera-motion controller.
Resolution and sound
Section titled “Resolution and sound”Choose 360p, 720p, 1080p or 4K for generation or editing. New nodes start at 360p with native sound. Existing nodes with an explicit resolution keep it; an older node without a resolution value retains its 720p fallback.
Generation accepts 3–10 seconds. Editing follows the source length. The connected API always generates native sound and offers no sound-off control. Review the current quote when changing resolution or length.
Settings were checked against the Omni generation API and edit API on 20 October 2026. No additional rendered result is claimed for the newly selected defaults.
| Setting | Value |
|---|---|
| Model ID | gemini_omni_flash |
| Name | Gemini Omni Flash |
| Availability | available |
| Default for new nodes | no |
Common declared limits
Section titled “Common declared limits”| Setting | Value |
|---|---|
| Capabilities | text to video, image to video, dialogue, reference video |
| Maximum prompt length | 20000 |
| Duration | 3–10 seconds; fallback 8 seconds |
| Audio | model-generated audio; fallback model-generated audio |
| Aspect ratio | 16:9, 9:16; fallback 9:16 |
| Declared language support | prompt: English; speech: English, Russian; tier: platinum |
| Combined reference limits | images: 10; videos: 1; audio inputs: 0; combined maximum: 10; audio without visual input: no; video reference seconds: minimum: 3; maximum: 10 |
| Controls | resolution: 360p, 720p, 1080p, 4k; new node 360p; fallback 720p |
| Element input form | sheet |
| Content constraints | people: restriction level: ok; minors: restriction level: risk; applies to: first frame, reference image, reference video |
| Negative prompt strategy | instruction in prompt text |
Mode: from_first_frame
Section titled “Mode: from_first_frame”| Setting | Value |
|---|---|
| First frame | required |
| Last frame | accepted; optional |
Mode: text_to_video
Section titled “Mode: text_to_video”Uses the common declared limits above.
Mode: from_references
Section titled “Mode: from_references”| Setting | Value |
|---|---|
| Combined reference limits | images: 10; videos: 3; audio inputs: 0; combined maximum: 13; audio without visual input: no; video reference seconds: minimum: not declared; maximum: 3; trim: head; videos reliable: 1; videos reliable source: Google |
| Video reference | light |
| Source video | accepted; optional; explicit mode selection required; scene clip is not inserted automatically |
| References | maximum: 10 |
Mode: edit_video
Section titled “Mode: edit_video”| Setting | Value |
|---|---|
| Duration | same as source video; fallback same as source video |
| Aspect ratio | not declared |
| Source video | required |
Unavailable modes: avatar_audio, extend_video, motion_transfer, recast.
Limits apply together. initial is the studio choice for a new node; default is the fallback for an existing node without that value. Declared language or input support does not guarantee a successful result.
Prompting playbook
Section titled “Prompting playbook”- Describe the environment, visible action, camera and lighting in coherent prose.
- For an edit, name the one change and list the properties of the source that must stay.
- When combining a character sheet and a clip, distinguish identity from the clip’s scene and movement.
The built-in Gemini Omni Flash authoring skill adapts a scene request for this family through Prompt for the model. It prepares instructions; it does not render video or certify the result.
Studio example
Section titled “Studio example”Select Gemini Omni Flash in edit mode with an approved daytime street clip. Ask Prompt for the model: “Make the pavement wet and the sky overcast. Keep the same person, coat, walking direction and camera path.” Inspect both the intended change and the preserved parts.
A run may consume credits. Review the prompt and inputs, then inspect the finished clip for action, identity, framing and sound. For automation, use the supported MCP workflow.
Known failure modes
Section titled “Known failure modes”The 5 October weather edit changed the street successfully, but rain was faint and evening looked more like an overcast day. A style sample transferred warmth and backlight, while its overhead and low camera movements did not transfer exactly.
Evidence and limits
Section titled “Evidence and limits”On 5 October 2026, the character-sheet reference test retained the character without showing sheet panels. A sheet plus character clip also passed, but that clip used the same scene; transfer to an unrelated scene remains untested. A style-clip test was partial for camera adherence. These results concern the v1.1 research entry, not older Omni releases.
The studio registry establishes available inputs and modes. A recorded render establishes only its observed outcome. The applicability guide explains the distinction. See also character identity, static camera and video style samples.
See the model–technique applicability matrix for connected operations, dated outcomes and configurations that remain untested.
Reviewed 14 October 2026 against the current studio contract and dated studio research. Official model-author documentation checked on that date; author capabilities may exceed Vini’s connected controls. Recheck this playbook when the model version, quality or mode changes.