Seedance 2.5 API · BytePlus
Dreamina Seedance 2.5 is ByteDance’s multimodal video model, supplied through BytePlus for this dreamina ID. BytePlus documents generation, editing and extension up to 30 seconds, with 480p, 720p, 1080p, 4k output. Check this entry’s current pricing, access and API route on XiuAI.[1][4]
dreamina-seedance-2-5-260628Model overview
Published provider specifications. Available features depend on the current API route.
- Input
- Text, image URLs, video URLs and audio URLs
- Output
- Video with optional synchronized audio
- Duration
- 4–30 seconds
- Resolution
- 480p / 720p / 1080p / 4K
- Reference limits
- 30 images · 10 videos · 10 audio clips
- Frame rate
- 24 fps
- API routes
- API access unavailable
For longer clips or more references
For a shot with several actions, product changes or audio cues, describe the stages first and give each reference a clear purpose. BytePlus’s 2.5 specification permits up to 30 images, 10 videos and 10 audio clips. These are limits, not quality guarantees; remove references that conflict with one another.[1]
First-frame, first/last-frame and omni-reference generation are separate modes with different reference roles. The provider editing mode requires reference video, an adaptive aspect ratio and automatic duration. Check the task mode when preserving a source clip rather than copying duration settings from ordinary generation.[1]
Choose resolution for the delivery environment
BytePlus documents 4K for 2.5, so a resolution limit from another supplier is not interchangeable. Its 1080p and 4K output uses 10-bit color and H.265/HEVC; check playback and editing compatibility before delivery. More pixels also change token usage, so one rate alone does not determine total cost.[1][2]
Understand video-token charges
This catalog entry is priced by video-token usage and the applicable rate. Resolution, reference video and actual output affect the charge; video inputs may also have minimum billable usage. Match input conditions and output specifications before comparing total costs across versions, rather than comparing only the rate per million tokens.[2][4]
The page reads rates and cost examples from the current catalog. Each example has its own resolution, duration and input assumptions; it is not a universal per-second quote. Checking a task and creating another generation are different actions. If submission is unclear, recover the original task before starting another generation.[4][2][3]
API access
Confirm access on XiuAI
This BytePlus entry uses the exact ID dreamina-seedance-2-5-260628. It has its own supplier, specifications and pricing, separate from the Volcengine entry with a doubao prefix. Replacing the prefix alone is not a migration. Check access with your key and use an API route actually listed on the page. If no video API or playground entry is shown, confirm the route with support before integrating.[4][5]
The BytePlus API creates asynchronous video tasks: save the returned task ID, query that task and retrieve its result. Video-task fields are different from chat messages; an OpenAI-compatible catalog label is not a reason to send the model to a chat endpoint. Provider parameters describe model capabilities, not a guarantee that every parameter is exposed by the current XiuAI integration.[1][3][5]
See the API documentation for the protocol and request fields supported by this model.
Read the API documentationView task and charge recordsTasks and data handling
- Asynchronous result
- Query by task ID after submission. Closing the page does not cancel generation.
- Result download
- Use the expiry returned with the task and save the result to your own storage before it expires.
- Input data
- Use accessible URLs you manage. XiuAI does not host input assets or manage creative workflows.
Frequently asked questions
How should I choose between 2.5 and Flash?
Start with the deliverable: 2.5 offers 480p / 720p / 1080p / 4k, while Flash offers 480p / 720p, with different reference limits. If Flash fits the required specification, compare current total costs and your own results under matching inputs and duration; its name is not a speed guarantee.[1][4]
Can I use audio references without images or video?
BytePlus’s 2.5 series permits audio-only input in omni-reference tasks; the 2.0 series does not. Audio references guide the task, while generate_audio controls whether output sound is requested. Check the two separately.[1]
Is Draft a free preview?
No. BytePlus fixes 2.5 Draft output at 480p and bills its usage and rate like normal 480p video. Generating a final video from a Draft is a separate task relationship. Check whether your current route exposes it; the provider feature is not a free XiuAI preview.[2][5]
Do I need a separate BytePlus resource pack for XiuAI?
Activation balances, resource packs and plans on BytePlus apply to accounts using that platform directly. Check XiuAI access, service group and charges against your XiuAI key, account balance and current prices. A provider resource-pack allowance is not an XiuAI allowance.[1][4][5]
Does task acceptance confirm the final video and cost?
No. Query the task after acceptance, save the result to your own storage when it is available, and reconcile usage and charges in task and request records. A temporary query failure does not establish generation failure, and a missing charge should not be treated as zero.[3][4]
Related models
Seedance 2.5 Flash · BytePlusSeedance 2.0 · BytePlusSeedance 2.5 · VolcengineSources and updates
Specifications reference provider documentation; prices and availability come from the current XiuAI catalog. Specifications checked: 2026-11-16
Content reviewed: 2026-11-16
[1] BytePlus video task parameters[2] BytePlus model pricing and billing[3] BytePlus video task retrieval[4] Current XiuAI model and pricing data[5] XiuAI protocol and access boundaries