Unlimited Open Source Models

Get Plan
Skip to main content
Available now on ModelsLab · ai

Seedance 2.0 API — ByteDance Video GenerationCinematic video + native audio via REST API. No BytePlus account.

Try Seedance 2.0API Documentation

How much does the Seedance 2.0 API cost per second?

On ModelsLab, Seedance 2.0 text to video costs $0.10 per second of video at 480p, $0.23 at 720p, $0.45 at 1080p and $0.94 at 4K, so a 5-second 720p clip is $1.15. Seedance 2.0 Fast is $0.08 per second at 480p and $0.18 at 720p, and Seedance 2.0 Mini is $0.044 and $0.0924. The price is the same with or without generated audio, and a Seedance 2.0 clip can be 4 to 15 seconds long.

Seedance runs on ByteDance’s BytePlus, so it is not part of the $149 Open Source Unlimited plan. Each clip is billed from the usage included in a ModelsLab plan (from $21/month), then from wallet balance, with no BytePlus account.

Some providers are cheaper. On September 27, 2026, Replicate listed Seedance 2.0 at $0.08 per second at 480p and $0.18 at 720p, and Kie.ai at $0.095 and $0.205, both below ModelsLab. At 1080p Replicate is the same $0.45. fal.ai ($0.3034 per second at 720p) and WaveSpeed ($0.24) cost more than ModelsLab at 720p.

Seedance API Price per Second on ModelsLab

Modelmodel_id480p720p1080p5 s at 720p
Seedance 2.0seedance-2-text-to-video$0.10$0.23$0.45$1.15
Seedance 2.0 Fastseedance-2.0-fast-t2v$0.08$0.18Not offered$0.90
Seedance 2.0 Miniseedance-2-mini-t2v$0.044$0.0924Not offered$0.462
Seedance 2.0 Multi-Referenceseedance-2-multi-reference-video$0.20$0.46$2.34$2.30
Seedance 2.5seedance-2-5-text-to-video$0.1236$0.2781Not offered$1.39
Seedance 1.5 Proseedance-1-5-pro$0.03$0.0624$0.1392$0.312

Prices checked September 27, 2026 from each model’s price rule: a base price per second at 480p times a resolution multiplier (for Seedance 2.0: 720p ×2.3, 1080p ×4.5, 4K ×9.4). Billed per second of output video.

Seedance 2.0 Price per Second by Provider

Provider480p720p1080pNotes
ModelsLab$0.10$0.23$0.45Same price with or without audio
Replicate$0.08$0.18$0.45Without a video input; $0.10, $0.22 and $0.55 with one
Kie.ai$0.095$0.205$0.51Without a video input
WaveSpeed$0.12$0.24$0.60Without a reference video
fal.aiAbout $0.134$0.3034$0.682With audio; billed per video token, 480p worked out from the token rate

Prices checked September 27, 2026 on replicate.com/bytedance/seedance-2.0, kie.ai/seedance-2-0, wavespeed.ai/models/bytedance/seedance-2.0/text-to-video and fal.ai/models/bytedance/seedance-2.0/text-to-video. ByteDance’s own BytePlus ModelArk bills per million video tokens; its pay-as-you-go rate is not in this table.

Seedance 2.0 — built for production video workflows

Seedance 2.0

ByteDance multimodal video model

Generate cinematic AI videos with Seedance 2.0 — ByteDance's flagship multimodal video generation model. Supports text, image, audio, and video as references.

Multi-Reference

Combine images, audio, and video inputs

Use multiple reference images and clips to guide subject, style, and motion in a single call. Available at /models/byteplus/seedance-20-multi-reference-to-video.

Start/End Frame Control

Precise narrative control

Pin the opening and closing frames of a clip. Seedance 2.0 generates the motion between them — perfect for animated transitions and narrative beats. Available at /models/byteplus/seedance-20-start-end-frame-image-to-video.

Native Audio Sync

Audio + video in one call

First industry model with unified audio-video joint generation. Sound effects, ambient audio, and dialogue timing aligned to the generated footage.

One REST API

No BytePlus account required

Call Seedance via the same ModelsLab REST API you use for Kling, Wan, and Runway Aleph. One key, one bill, no BytePlus signup or Volcano Engine quota management.

Pricing

From $0.044 per second

Seedance 2.0 Mini starts at $0.044 per second of 480p output; Seedance 2.0 at 720p is $0.23 per second (checked 2026-09-23). Billed from the usage included in your plan, then from wallet balance.

Webhooks

Async callbacks for long renders

Pass a webhook URL and ModelsLab POSTs the finished MP4 to your endpoint when generation completes. Essential for batch and production pipelines.

Python & JavaScript

Full SDK coverage

Official Python and TypeScript SDKs wrap the Seedance endpoints. REST + OpenAPI spec for autogenerated clients in any language.

Examples

See what Seedance 2.0 can create

Copy any prompt below and try it yourself in the playground.

Cinematic product reveal

“a sleek smartphone rotating on a minimalist white pedestal, studio lighting, shallow depth of field, cinematic, 1080p”

Multi-reference character scene

“the character from reference image 1 walking through the environment from reference image 2, golden hour lighting”

Start/end frame animation

“morph from the frame-1 pose to the frame-2 pose with natural physics and fluid motion”

Native audio video

“a waterfall crashing into a pool with natural ambient sound and reverberant splash audio”

For Developers

A few lines of code.
One endpoint. Multi-reference input. Cinematic output.

ModelsLab handles the infrastructure: fast inference, auto-scaling, and a developer-friendly API. No GPU management needed.

  • Serverless: scales to zero, scales to millions
  • Pay per second, on plans from $21/month
  • Python and JavaScript SDKs, plus REST API
API Documentation 
import requests


response = requests.post(
    "https://modelslab.com/api/v7/video-fusion/text-to-video",
    json={
  "key": "YOUR_API_KEY",
  "prompt": "a dancer performing fluid contemporary moves in a sunlit studio, cinematic camera pan",
  "duration": 5,
  "model_id": "seedance-2-text-to-video",
  "resolution": "720p",
  "aspect_ratio": "16:9",
  "generate_audio": true
}
)
print(response.json())

FAQ

Common questions about Seedance 2.0 API — ByteDance Video Generation

Read the docs

Seedance 2.0 is ByteDance's flagship multimodal video generation API. It accepts text, image, audio, and video inputs and produces cinematic 1080p video clips with synchronized native audio. On ModelsLab you access Seedance 2.0 via a single REST endpoint — no BytePlus or Volcano Engine account required.

Seedance 2.0 is billed per second of output video, scaled by resolution: $0.10 per second at 480p, $0.23 at 720p, $0.45 at 1080p and $0.94 at 4K, so a 5-second 720p clip is $1.15. Seedance 2.0 Fast is $0.08 and $0.18 per second at 480p and 720p, and Seedance 2.0 Mini $0.044 and $0.0924. Generated audio does not change the price. It is a third-party model, billed from the usage included in a ModelsLab plan from $21/month, then from wallet balance. Prices checked 2026-09-27.

ByteDance sells Seedance 2.0 through BytePlus ModelArk, and resellers include Replicate, fal.ai, Kie.ai, WaveSpeed and ModelsLab. On 2026-09-27 Replicate was the cheapest at 720p ($0.18 per second without a video input), then Kie.ai ($0.205), ModelsLab ($0.23), WaveSpeed ($0.24) and fal.ai ($0.3034). At 1080p Replicate and ModelsLab both list $0.45 per second. ModelsLab puts Seedance on the same key and plan as Kling, Veo, Wan and 10,000+ other models, with no BytePlus account.

No. Seedance runs on BytePlus, so every call is billed from the usage included in a ModelsLab plan from $21/month, then from wallet balance. Open Source Unlimited covers only the open-weight models ModelsLab self-hosts.

Create a ModelsLab account at modelslab.com, subscribe to a plan (from $21/month), and generate your API key from the dashboard. The same key works for every Seedance variant (1.0 Pro, 1.5 Pro, 2.0) plus every other video model (Kling, Wan, Runway Aleph).

Endpoint reference, parameter schemas, request/response examples, and SDK snippets live at docs.modelslab.com, and the price table at the top of this page compares every Seedance version. Each Seedance variant has its own model page at /models/byteplus/ with direct-copy code samples.

Seedance 2.0 adds multimodal input (text + image + audio + video references in a single call), native audio-video joint generation, and tighter motion coherence across longer clips. Seedance 1.5 Pro is text/image-to-video only and cheaper per clip, which makes it the right pick for high-volume basic workflows.

Yes. The "start/end frame" variant takes a first-frame and last-frame image and animates the transition between them. The "multi-reference" variant accepts multiple image references to control subject, style, and scene.

Seedance 2.0 takes a duration of 4 to 15 seconds per call, and Seedance 2.5 up to 30. For longer sequences, chain multiple calls at scene boundaries and stitch the clips together — a common pattern for 30-60s social videos.

Yes. Pass a webhook URL in the request body and ModelsLab POSTs the completed MP4 URL to your endpoint when generation finishes. Essential for long-running generations and high-throughput batch jobs.

Seedance 2.0 wins on multimodal input and native audio. Kling 3.0 leads cinematic motion realism. Runway Aleph is purpose-built for video-to-video editing (not generation). Wan 2.7 is cheapest for bulk text-to-video. All four are available under one ModelsLab API key — route between them per use case.

The official ModelsLab Python and TypeScript SDKs cover every Seedance endpoint. Community GitHub wrappers also exist (Anil-matcha/Seedance-2) but the official SDK is recommended for production — it handles retries, webhooks, and async polling out of the box.

Ready to create?

Start generating with Seedance 2.0 API — ByteDance Video Generation on ModelsLab.

Try Seedance 2.0API Documentation