Wan
Wan is a video model family by Alibaba. Wan offers Base and Prime choices in Vini. Use an explicit scene plan for longer video and inspect each shot. A video sample may carry its content into a result, so attaching it as “style only” is not enough to guarantee separation.
Resolution, quality and sound
Section titled “Resolution, quality and sound”Generation offers Base or Prime at 480p, 720p or 1080p. New nodes start at Base, 480p, with sound on. Existing explicit settings remain selected; an older node without a resolution value retains its 720p fallback. Turning generated sound off is supported and does not lower the provider’s listed rate.
Generation accepts 2–30 seconds. Editing has a different set of controls: 720p or 1080p, no quality choice, the source length and the source sound. It has no sound-off switch. The editing minimum therefore remains 720p even when a new generation starts at 480p.
Settings were checked against the Wan 3.0 API, Prime API and editing API on 20 October 2026. This check does not transfer Base’s recorded outcomes to Prime.
| Setting | Value |
|---|---|
| Model ID | wan |
| Name | Wan |
| Availability | available |
| Default for new nodes | no |
Common declared limits
Section titled “Common declared limits”| Setting | Value |
|---|---|
| Capabilities | text to video, image to video, native audio, dialogue, reference video, reference audio |
| Prompt required | yes |
| Maximum prompt length | 20000 |
| Duration | 2–30 seconds; fallback 5 seconds |
| Audio | model-generated audio, off; fallback model-generated audio |
| Aspect ratio | 9:16, 16:9, 1:1, 4:3, 3:4; fallback 9:16 |
| Declared language support | prompt: English; speech: English; tier: platinum |
| Combined reference limits | images: 10; videos: 5; audio inputs: 5; combined maximum: 20; audio without visual input: yes; video reference seconds: minimum: 1; maximum: 15; combined maximum: 15; audio reference seconds: minimum: 1; maximum: 15; combined maximum: 15; trim: head; video fps min: 16 |
| Controls | quality: base, prime; fallback base; resolution: 480p, 720p, 1080p; new node 480p; fallback 720p |
| Element input form | sheet |
| Content constraints | people: restriction level: unknown |
| Shots | seconds per shot: 5; shot timing format: timecodes |
| Negative prompt strategy | instruction in prompt text |
Mode: from_first_frame
Section titled “Mode: from_first_frame”| Setting | Value |
|---|---|
| Prompt required | no |
| 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 |
|---|---|
| Prompt required | no |
| Video reference | content |
| Source video | accepted; optional |
| References | maximum: 10 |
Mode: edit_video
Section titled “Mode: edit_video”| Setting | Value |
|---|---|
| Maximum prompt length | 5000 |
| Duration | same as source video; fallback same as source video |
| Audio | origin; fallback origin |
| Aspect ratio | same as source video |
| Combined reference limits | images: 1; videos: 1; audio inputs: 0; combined maximum: 2; audio without visual input: no; video reference seconds: minimum: 2; maximum: 10; combined maximum: 15; audio reference seconds: minimum: 1; maximum: 15; combined maximum: 15; trim: head; video fps min: not declared |
| Controls | quality: not declared; resolution: 720p, 1080p; new node 720p; fallback 720p |
| Shots | not available in this configuration |
| Source video | required |
| References | maximum: 1 |
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”- Give each shot a clear action, framing and transition. Keep actions in chronological order.
- Translate a sample into descriptions of camera, light and rhythm when its people and place must not appear.
- Keep reference roles and subject identities stable across shots. Do not treat Base evidence as Prime evidence.
The built-in Wan 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”Choose Wan Base with character and place references. Ask Prompt for the model: “First show the closed greenhouse from outside. Then a medium shot as the gardener opens the door. Finally follow the gardener one step inside.” Review continuity between the three shots.
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”In the 4 October Base video-style test, the final shot copied people and landscape from the sample. A later words-only instruction avoided that content, but camera timing and warmth still differed. Prime was not tested in those cells.
Evidence and limits
Section titled “Evidence and limits”The older reference and editing exams used individual frames. Their current sheet inputs have outstanding replacement exams; those historical passes are not current-input passes.
The 28 September 2026 reference and replacement-edit cases retained the fictional character. The reference request used Wan 3.0; the edit request used Wan 2.7. Those are different mode-specific version scopes. On 4 October, a words-only filming description produced the requested camera order without importing sample subjects, with deviations in duration and light. One success does not establish dependable long-form continuity.
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.