Seedance 2.0
Seedance 2.0 is a video model family by ByteDance. Choose a first frame when the opening composition matters, references when appearance matters, and edit or extension for an existing clip.
Seedance 2.0 quality: new and saved nodes
Section titled “Seedance 2.0 quality: new and saved nodes”Seedance 2.0 is one model family with Draft, Fast and Full quality choices. A new Video node uses Draft at 480p with sound on. An existing or saved Seedance 2.0 node without a quality value falls back to Full. An explicitly saved quality remains the selected value; the new-node default does not replace it.
Draft and Fast support 480p or 720p. Full additionally supports 1080p and 4K. Resolution and quality are separate choices; an unsupported higher resolution is not offered for Draft or Fast. Generation accepts 4–15 seconds, and sound can be turned off. Editing keeps the source length. An older node without a resolution value retains its 480p fallback.
Settings were checked against the BytePlus Seedance 2.0 documentation and pricing on 20 October 2026. Higher resolutions have different token costs; review the quote for the selected quality, resolution, length and references. These settings checks do not establish new rendered outcomes.
Keep outcome evidence separate for Draft, Fast and Full. A successful Fast render does not establish a successful Full or Draft result, even when the modes accept the same inputs.
Before a paid MCP run, use describe_studio → model_advice → price_quote: inspect the current catalog, read the video-model advice, then quote the exact model, mode, quality and inputs you intend to run. Review the estimate, warnings and spending ceiling before proceeding. The model-selection guide explains how to quote an individual operation or a saved graph node.
| Setting | Value |
|---|---|
| Model ID | seedance2 |
| Name | Seedance 2.0 |
| Availability | available |
| Default for new nodes | yes |
| Earlier IDs resolving to this record | seedance2_fast, seedance2_mini |
Common declared limits
Section titled “Common declared limits”| Setting | Value |
|---|---|
| Capabilities | text to video, image to video, native audio, dialogue, reference video, reference audio |
| Maximum prompt length | no fixed maximum declared |
| Duration | 4–15 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, 21:9; fallback 9:16 |
| Declared language support | prompt: English; speech: English, Chinese |
| Combined reference limits | video frame: min pixels: 407696; images: 9; videos: 3; audio inputs: 3; combined maximum: 12; audio without visual input: no; video reference seconds: minimum: 2; maximum: 15; combined maximum: 15; audio reference seconds: minimum: 2; maximum: 15; combined maximum: 15; trim: head |
| Controls | quality: draft, fast, full; new node draft; fallback full; resolution: 480p, 720p; new node 480p; fallback 480p |
| Element input form | frames |
| Content constraints | people: 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 |
Quality: full in from_first_frame
Section titled “Quality: full in from_first_frame”| Setting | Value |
|---|---|
| Controls | quality: draft, fast, full; new node draft; fallback full; resolution: 480p, 720p, 1080p, 4k; new node 480p; fallback 480p |
Mode: text_to_video
Section titled “Mode: text_to_video”Uses the common declared limits above.
Quality: full in text_to_video
Section titled “Quality: full in text_to_video”| Setting | Value |
|---|---|
| Controls | quality: draft, fast, full; new node draft; fallback full; resolution: 480p, 720p, 1080p, 4k; new node 480p; fallback 480p |
Mode: from_references
Section titled “Mode: from_references”| Setting | Value |
|---|---|
| Video reference | style |
| Source video | accepted; optional |
| References | maximum: 9 |
Quality: full in from_references
Section titled “Quality: full in from_references”| Setting | Value |
|---|---|
| Controls | quality: draft, fast, full; new node draft; fallback full; resolution: 480p, 720p, 1080p, 4k; new node 480p; fallback 480p |
Mode: edit_video
Section titled “Mode: edit_video”| Setting | Value |
|---|---|
| Duration | same as source video; fallback same as source video |
| Aspect ratio | same as source video |
| Source video | required |
| References | maximum: 9 |
Quality: full in edit_video
Section titled “Quality: full in edit_video”| Setting | Value |
|---|---|
| Controls | quality: draft, fast, full; new node draft; fallback full; resolution: 480p, 720p, 1080p, 4k; new node 480p; fallback 480p |
Mode: extend_video
Section titled “Mode: extend_video”| Setting | Value |
|---|---|
| Aspect ratio | same as source video |
| Source video | required |
| References | maximum: 9 |
Quality: full in extend_video
Section titled “Quality: full in extend_video”| Setting | Value |
|---|---|
| Controls | quality: draft, fast, full; new node draft; fallback full; resolution: 480p, 720p, 1080p, 4k; new node 480p; fallback 480p |
Unavailable modes: avatar_audio, 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 scene as an ordered sequence of visible actions, with a clear starting state.
- Give every reference a role: character, location, movement or sound. A reference image is not automatically the opening frame.
- Treat a style clip as optional evidence about filming. Describe the intended camera and light in words and keep the subjects of the sample out of the brief.
The built-in Seedance 2.0 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”Prepare clean frames of a fictional gardener as an Element. Choose Seedance 2.0, Fast, and reference mode. Ask Prompt for the model: “The gardener stands beside an empty pot, lowers a seedling into it, then presses the soil with both hands. One steady medium shot.” Check the generated prompt and input roles before running Video.
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”A character sheet was refused by input moderation in an extension test on 30 September; clean frames of the same fictional character passed. This does not promise acceptance of every photograph. The tested style-reference result began close to the location image, so a location reference can unintentionally anchor the opening angle.
Evidence and limits
Section titled “Evidence and limits”Fast reference generation and editing retained the tested character on 29 September 2026. Fast extension retained the character and setting on 30 September. On 4 October, Fast and Full style-reference cases carried lighting and cutting rhythm without sample subjects; Draft was partial. These observations are separate by quality. No dedicated static-camera success claim is made.
A reference sound longer than 15 seconds is not refused: Vini sends its first 15 seconds and says so on the Video card before the run. The 15-second total counts what is sent, so two 16-second songs are still refused. The cut is made by the studio; no render was made to judge where it falls in the music.
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.
Studio guidance reviewed 15 October 2026 against the current studio contract and existing dated research. Official model-author documentation checked on 14 October 2026; author capabilities may exceed Vini’s connected controls. Recheck this playbook when the model version, quality or mode changes.