Skip to main content
ByteDanceVideo model

Seedance 2.5

ByteDance's newest video model, built to tell a whole scene in one 30-second take, with sound.

Build with Seedance 2.5

Each clip is one prompt, run on Seedance 2.5 in Runable.

Seedance 2.5 on Runable

  • 30-second takes

    A full scene, with setup and payoff, in one pass.

  • Direct with references

    Add up to 30 images, 10 videos and 10 audio clips.

  • No ByteDance account

    Runs on your Runable credits, next to other video models.

How to use Seedance 2.5 on Runable

  1. Select Seedance 2.5

    Start a task on Runable and pick Seedance 2.5 in the model menu.

  2. Describe the shot

    Type the scene, or add images and clips for it to follow.

  3. Download the video

    Refine by chatting, then download the clip or cut it into an edit.

One prompt, finished work

Real results from single Runable tasks. Pick one to watch it.

Made on Runable with Seedance 2.5
One unbroken handheld shot. A jazz singer in a green velvet dress pushes through a curtain onto a small club stage and starts singing at the microphone as the audience applauds. Warm tungsten light, haze in the air. Realistic film look, natural skin texture.
longest clip in one pass
30s
references in one generation
50
output, with sound
1080p
turn to build
Your

Why choose Seedance 2.5

  • Twice the length

    30 seconds in one pass, double Seedance 2.0’s 15.

  • A story in one take

    ByteDance says it links several shots into one continuous scene.

  • Sound with the picture

    Dialogue, effects and music are generated together, in sync.

  • The most references

    Up to 30 images, 10 videos and 10 audio clips per shot.

  • Less of the AI look

    ByteDance tuned skin, eyes and lighting to look more like real footage.

  • First and last frames

    Animate a keyframe, with an optional end frame.

Seedance 2.5 specs

Maker
ByteDance (Seed team)
Released
July 31, 2026
Clip length
4 to 30 seconds in one pass
Resolution
480p, 720p or 1080p
Aspect ratios
21:9, 16:9, 4:3, 1:1, 3:4, 9:16; image-to-video follows your image
Inputs
Text, a start frame with an optional end frame, or references
References
Up to 30 images, 10 videos and 10 audio clips
Audio
Native, on by default
Specs are for Seedance 2.5 on Runable. Reference videos, and reference audio, can run up to 30 seconds in total. ByteDance’s own apps add multi-round extension and timestamp-level editing. On Runable you use your plan’s credits instead of a Dreamina or Jimeng subscription.

The only AI agent you need

Compare plansHelp

Need SSO, data localization, premium support, onboarding assistance? Contact Sales

Seedance 2.5 FAQs

Seedance 2.5 is a video generation model from ByteDance’s Seed team. It launched on July 31, 2026, and makes clips of up to 30 seconds with sound in one pass, taking direction from images, videos and audio you provide.

Sources: ByteDance Seed: Introducing Seedance 2.5 · ByteDance Seed: Seedance 2.5 · fal: Seedance 2.5 text to video · Runable: Seedance 2.5 is live on Runable

More AI models on Runable

Put Seedance 2.5 to work

Describe the job. Runable runs it on Seedance 2.5 and ships the result.