Seedance 2.5
Seedance 2.5 generates 30-second clips at native 4K with synchronized audio, accepting up to 50 multimodal reference inputs spanning images, video, audio, and style references.
View API reference- Output price
- Output $10.70, Per 1M generated tokens
import { experimental_generateVideo as generateVideo } from 'ai';
const result = await generateVideo({ model: 'bytedance/seedance-2.5', prompt: 'A serene mountain lake at sunrise.'});Copy link to headingPlayground
Try out Seedance 2.5 by ByteDance. Usage is billed to your team at API rates. Free users (those who haven't made a payment) get $5 of credits every 30 days.
Seedance 2.5
Your generated video will appear here.
Copy link to headingProviders
Route requests across multiple providers. Copy a provider slug to set your preference. Visit the docs for more info. Using a provider means you agree to their terms, listed under Legal.
Getting started
Generate videos with Seedance 2.5 using the experimental_generateVideo function from AI SDK 6 or later. AI Gateway handles routing and polls until the video is ready.
Install the AI SDK (pnpm add ai dotenv), create an API key from the API Keys page, and set it as AI_GATEWAY_API_KEY in your environment. Full setup is covered in the video generation quickstart.
import { experimental_generateVideo as generateVideo } from 'ai';import fs from 'node:fs';import 'dotenv/config';
async function main() { const result = await generateVideo({ model: 'bytedance/seedance-2.5', prompt: 'A chicken flying into the sunset in the style of 90s anime', });
// Save the generated video fs.writeFileSync('output.mp4', result.videos[0].uint8Array);
console.log('Video saved to output.mp4');}
main().catch(console.error);Top-level parameters
Exercise the supported top-level params: prompt, aspectRatio, resolution, and duration.
import { experimental_generateVideo as generateVideo } from 'ai';import fs from 'node:fs';import 'dotenv/config';
async function main() { const result = await generateVideo({ model: 'bytedance/seedance-2.5', prompt: 'A chicken flying into the sunset in the style of 90s anime', aspectRatio: '16:9', resolution: '1280x720', duration: 5, });
// Save the generated video fs.writeFileSync('output.mp4', result.videos[0].uint8Array);
console.log('Video saved to output.mp4');}
main().catch(console.error);| Parameter | Type | Required | Description |
|---|---|---|---|
prompt | string | No | Text description of the video to generate. |
duration | number | No | Video length in seconds. 4-30 seconds. |
resolution | string | No | Resolution ('854x480', '1280x720', '1920x1080'). |
aspectRatio | string | No | Aspect ratio ('16:9', '4:3', '1:1', '3:4', '9:16', '21:9'). |
generateAudio | boolean | No | Generate synchronized audio with the video. |
frameImages | Array<{ image: string; frameType: 'first_frame' | 'last_frame' }> | No | First and last frames of the clip. A first_frame entry replaces prompt.image and wins when both are set, and adding a last_frame transitions between the two. Seedance accepts image URLs only, so host local files on Vercel Blob first. |
inputReferences | Array<{ data: string; mediaType: string }> | No | Reference images and videos, referenced in the prompt as [Image 1], [Video 1], and so on, numbered separately in the order you pass them. Tag every URL with an explicit mediaType — an untyped URL is treated as an image and emits a warning. See the Input limits table for supported counts. |
Input limits
| Input | Formats | Sources | Max count | Max size | Limits |
|---|---|---|---|---|---|
| Image | jpeg, png, webp, bmp, tiff, gif, heic, heif | url | 30 | 30 MB | ≥300px · ≤6000px · aspect 2:5–5:2 |
| Video | mp4, mov | url | 10 | 200 MB | 2-30s · ≥300px · ≤6000px |
| Audio | mp3, wav | url | — | 15 MB | 2-30s |
Provider options (bytedance)
Load the compatible Seedance options under providerOptions.bytedance. Frames and references are passed at the top level through frameImages and inputReferences, which change the call shape and are shown in their own examples below.
import { experimental_generateVideo as generateVideo } from 'ai';import fs from 'node:fs';import 'dotenv/config';
async function main() { const result = await generateVideo({ model: 'bytedance/seedance-2.5', prompt: 'A chicken flying into the sunset in the style of 90s anime', resolution: '1280x720', duration: 5, providerOptions: { bytedance: { seed: 42, pollIntervalMs: 5000, pollTimeoutMs: 600000, }, }, });
// Save the generated video fs.writeFileSync('output.mp4', result.videos[0].uint8Array);
console.log('Video saved to output.mp4');}
main().catch(console.error);Pass these Seedance-specific options under providerOptions.bytedance in your generateVideo call.
| Parameter | Type | Required | Description |
|---|---|---|---|
lastFrameImage | string | No | URL of the last frame image, enabling first+last frame mode. Legacy alternative to the top-level frameImages, used only when frameImages is omitted. |
referenceImages | string[] | No | Reference image URLs for reference-to-video, referenced in the prompt as [Image 1], [Image 2], and so on — see the Input limits table for the supported count. Legacy alternative to the top-level inputReferences, used only when inputReferences is omitted. |
referenceVideos | string[] | No | Reference video URLs for reference-to-video, numbered separately from the images and referenced in the prompt as [Video 1], [Video 2], and so on. Legacy alternative to the top-level inputReferences, used only when inputReferences is omitted. |
referenceAudio | string[] | No | Reference audio URLs, sent alongside the reference images and videos to drive the generated audio. |
seed | number | No | Fix the random seed for reproducible output. |
pollIntervalMs | number | No | How often to check task status. Defaults to 3000. |
pollTimeoutMs | number | No | Maximum wait time. Defaults to 300000 (5 minutes). |
Frames take priority over references
Frames and references are mutually exclusive. When frameImages is set, inputReferences and the legacy providerOptions.bytedance.referenceImages / referenceVideos are dropped with a warning.
The top-level parameters win over their provider-option equivalents: frameImages overrides prompt.image and lastFrameImage, and inputReferences overrides referenceImages and referenceVideos. Set one or the other, not both.
providerOptions.bytedance.referenceAudio has no top-level equivalent, so it stays a provider option and is sent alongside whichever reference path you use.
Reference-to-video
Pass reference media through the top-level inputReferences so the model keeps subjects, style, and composition consistent — see the Input limits table for the supported counts. Reference each one in the prompt with [Image 1], [Video 1], and so on; images and videos are numbered separately in the order you pass them. Tag every URL with an explicit mediaType, since Seedance cannot infer image or video from a bare URL.
import { experimental_generateVideo as generateVideo } from 'ai';import fs from 'node:fs';import 'dotenv/config';
async function main() { const result = await generateVideo({ model: 'bytedance/seedance-2.5', prompt: 'The lion from [Image 1] walks into the clearing from [Video 1], settles in the long grass, and watches the sun go down.', aspectRatio: '16:9', resolution: '1280x720', duration: 30, generateAudio: true, inputReferences: [ { data: 'https://example.com/lion.jpg', mediaType: 'image/jpeg' }, { data: 'https://example.com/clearing.mp4', mediaType: 'video/mp4' }, ], providerOptions: { bytedance: { pollTimeoutMs: 600000, }, }, });
// Save the generated video fs.writeFileSync('output.mp4', result.videos[0].uint8Array);
console.log('Video saved to output.mp4');}
main().catch(console.error);Copy link to headingMore models by ByteDance
| Model |
|---|
| — | $0.003/M | $0.04/img | — | | 07/11/2026 | |||||||||
| — | $0.04/img | — | | 02/13/2026 | ||||||||||
| — | $0.04/img | — | | 12/03/2025 | ||||||||||
| — | $0.03/img | — | | 09/09/2025 | ||||||||||
| 256K | 1.1 s | 85 tps | $0.25/M+1 more | $2/M+1 more | Read$0.05/M | — | | 09/01/2025 | ||||||
| 256K | 1.1 s | 104 tps | $0.25/M+1 more | $2/M+1 more | Read$0.05/M | — | | 09/01/2025 |
Copy link to headingAbout Seedance 2.5
Seedance 2.5 is the generation after Seedance 2.0 in ByteDance's Seed family, skipping several version numbers to signal the size of the jump.
Two capabilities define it. Single-shot length reaches 30 seconds, with multi-turn extension for longer sequences. Resolution is native 4K rather than upscaled from a smaller render, which matters for professional pipelines where an upscale is visible, and 10-bit colour depth leaves headroom for grading in post.
The larger change is reference capacity. Seedance 2.5 accepts up to 50 multimodal inputs in a single request, spanning images, video, audio clips, 3D white models, and style references, against 12 in its predecessor. That is what makes it directable: style, motion, and composition can be shown rather than described. ByteDance also reports roughly 20 percent better prompt adherence, which reduces the number of generations needed to reach a usable result.
Audio and video are generated together by one model rather than dubbed afterwards, inheriting the joint architecture introduced with Seedance 2.0.
Generate video with generateVideo from the AI SDK.
Copy link to headingWhat To Consider When Choosing a Provider
- Configuration: Reference capacity is the reason to choose Seedance 2.5, and it is also the work. Fifty inputs give fine control over style, motion, and composition, but assembling and curating that reference set is a real production step. On a simple prompt-to-clip task the earlier Seedance models reach a usable result with less setup.
- Configuration: Native 4K and 30-second duration both raise cost per generation compared with shorter, smaller output. Prototype at lower resolution and duration, then generate the final at full settings. See the pricing panel on this page for current rates.
- Configuration: Regional availability has been staged. Confirm the model is reachable for your account and region before you design a pipeline around it.
- Zero Data Retention: Zero Data Retention is offered on a per-provider and model basis. See the documentation for details.
- Authentication: AI Gateway authenticates requests using an API key or OIDC token. You do not need to manage provider credentials directly.
Copy link to headingWhen to Use Seedance 2.5
Best for
- Native 4K Delivery: No visible upscale from a smaller render
- Reference-Heavy Direction: Up to 50 multimodal inputs in one request
- Longer Single Shots: 30 seconds before extension is needed
- Colour Grading Workflows: 10-bit depth leaving headroom in post
- Joint Audio And Video: Generated together rather than dubbed
Consider alternatives when
- Simple Prompt-To-Clip Work: Earlier Seedance models need less setup
- Cost-Constrained Generation: 4K and 30 seconds both raise cost per clip
- Short Clip Needs: A lite Seedance variant is cheaper
- Still Image Output: Seedream is the right family
Copy link to headingConclusion
Seedance 2.5 generates 30 seconds at native 4K and takes up to 50 reference inputs, which makes it a directable model rather than a prompt-and-hope one. Call bytedance/seedance-2.5 through generateVideo in the AI SDK, and budget for the reference curation that its control depends on.
Copy link to headingFrequently Asked Questions
How long can a Seedance 2.5 clip be?
Up to 30 seconds in a single generation, with multi-turn extension for longer sequences.
Is Seedance 2.5 really 4K?
Yes, generated natively at 4K rather than upscaled from a lower resolution render. It also supports 10-bit colour depth for grading in post.
How many reference inputs does Seedance 2.5 accept?
Up to 50 in one request, spanning images, video, audio clips, 3D white models, and style references. Its predecessor accepted 12.
How do I call Seedance 2.5 on AI Gateway?
Use
generateVideofrom the AI SDK withbytedance/seedance-2.5. The Chat Completions, Responses, and Messages APIs do not serve video generation.Does Seedance 2.5 generate audio?
Yes. Audio and video are produced together by one model, inheriting the joint generation architecture introduced with Seedance 2.0.
How does Seedance 2.5 compare to Seedance 2.0?
Longer single shots, native 4K, and a much larger reference budget of 50 inputs against 12. ByteDance also reports roughly 20 percent better prompt adherence.
Does Seedance 2.5 support Zero Data Retention?
Zero Data Retention is not currently available for this model. Zero Data Retention is offered on a per-provider basis. See https://vercel.com/docs/ai-gateway/capabilities/zdr for details.
Your use is subject to ByteDance's Terms & Privacy Policies.