50% off yearly plansLimited time — ends in 50% off yearly plans · ends in

Seedance 2.5 AdAnt · by ByteDance Seed

Seedance 2.5 AI Video Generator

Use Seedance 2.5 online: one prompt makes up to 30 seconds of video with sound, guided by images, videos and audio you add, up to 4K.

  • Up to 30 s
  • Native audio
  • Up to 4K
  • Image, video and audio references
Made with Seedance 2.5The same-prompt test: a UGC skincare ad, 6 s at 720p, one run, no edits.Watch video

Same prompt, every video model: a UGC skincare ad

We ran this prompt once on each model on AdAnt, with no retries and no edits: 9:16, 720p, 6 seconds, audio on. Whether she says the line clearly and in sync, whether the bottle keeps its shape in her hand, and whether it reads as a phone video rather than a film.

Handheld selfie-style UGC video. A woman in her late 20s with curly brown hair, in a sunlit bathroom, holds a small amber glass serum bottle up next to her face, smiles at the camera and says: "This is the only serum I've ever finished twice." Natural window light, real skin texture, slight handheld shake, vertical 9:16 phone framing.
Seedance 2.5Ready in 2 min 36 s · about 9 creditsWatch video
Gemini Omni FlashReady in 37 s · about 3 creditsWatch video
Seedance 2.0Ready in 4 min 7 s · about 6 creditsWatch video
Seedance 2.0 FastReady in 1 min 43 s · about 5 creditsWatch video
Topview T1Ready in 3 min 19 s · about 5 creditsWatch video

Video models compared

ModelMade byModesLongest clipResolutionAudioInputsTest runTest costBest for
Seedance 2.5ByteDance SeedText to video, image to video, reference to video30 s4KNativeStart/end image, up to 30 images, 10 videos, 10 audio clips2 min 36 sabout 9 creditslonger story clips and complex scenes with many references
Gemini Omni FlashGoogleText to video, image to video10 s4KNativeStart/end image, up to 30 reference images37 sabout 3 creditsfast drafts and talking clips with sound
Seedance 2.0ByteDance SeedText to video, image to video, video editing15 s4KNativeStart/end image, up to 9 images, 3 videos, 3 audio clips4 min 7 sabout 6 creditsUGC ads and product clips with references
Seedance 2.0 FastByteDance SeedText to video, image to video15 s4KNativeStart/end image, up to 9 images, 3 videos, 3 audio clips1 min 43 sabout 5 creditstesting many hooks and ideas cheaply
Topview T1TopviewText to video, image to video30 s1080pNativeStart image, optional end image3 min 19 sabout 5 creditslonger single takes and calm, natural clips

Test run and cost come from the same-prompt run above, measured on AdAnt. Longer clips and higher resolutions cost more; the try-it panel shows the price before you run.

Seedance 2.5 key features

Up to 30 seconds in one run

A full hook, demo and payoff in a single clip, without stitching, so the person, the light and the product stay the same throughout.

References of every kind

Guide a run with up to 30 images, 10 video clips and 10 audio clips: a face, a product, a camera move to copy, a voice to match.

Image to video with a locked first frame

Start from your own photo and Seedance 2.5 starts from that exact frame, with an optional last frame to land on.

Up to 4K, with sound

Render drafts at 480p or 720p and the final at 1080p or 4K, always with a generated audio track you can turn off.

  1. Describe the shot: who is in it, what happens, the line they say and the camera.
  2. Pick 9:16 for TikTok, Reels and Shorts, or 16:9 for YouTube.
  3. Press Create. The clip plays here with its sound when it is ready.

A Seedance 2.5 prompt for a longer clip

0–5 s: [the hook: a person, a problem, one line in quotes]. 5–15 s: [the product in use, one camera move]. 15–25 s: [the result, a second line]. 25–30 s: [the product on screen, a closing line]. Keep the same person and outfit throughout. [Light], [style].

Seedance 2.5: ByteDance's AI Video Model for 30-Second Clips with Audio

Seedance 2.5 is ByteDance Seed's video model that follows Seedance 2.0. It makes videos with native audio from a prompt and from image, video and audio references, on AdAnt up to 30 seconds in one run and up to 4K. On AdAnt it runs as text to video, reference to video and image to video.

Full short ads in one clip

A 15 to 30-second UGC ad with a hook, a demo and a call to action, made in one run.

Product demos

Show a product being used from more than one angle, with the reference keeping it the same.

Recreating a viral video

Pass the original as a video reference and Seedance 2.5 follows its pacing and camera.

Seedance 2.5 AI Video Generator FAQ

Tips for a better result
  • For anything over 10 seconds, write the prompt as timed beats.
  • Say "keep the same person and outfit throughout" in long clips.
  • Draft at 480p or 720p; render the winner at 1080p or 4K.
What is Seedance 2.5?

Seedance 2.5 is ByteDance Seed's video model after Seedance 2.0. It makes videos with sound from text, images, videos and audio, up to 30 seconds.

Is Seedance 2.5 free?

Running this model uses paid credits from an AdAnt plan. Creating an account is free. A 5-second 720p clip costs about 7 credits; longer and higher-resolution clips cost more.

Seedance 2.5 vs Seedance 2.0: what changed?

On AdAnt, Seedance 2.5 runs clips up to 30 seconds against 15 for Seedance 2.0, and takes far more references: 30 images, 10 videos and 10 audio clips against 9, 3 and 3. In our same-prompt test it was also faster, about 2.6 minutes against about 4, and cost about 9 credits against 6.

How long can a Seedance 2.5 video be?

From 4 to 30 seconds in one run on AdAnt.

Can Seedance 2.5 turn an image into a video?

Yes. Its image-to-video mode starts exactly from your image and can end on a second one.

Does Seedance 2.5 generate audio?

Yes, with the video. You can turn audio off when you will add a voiceover later.

Is there a Seedance 2.5 API?

ByteDance offers Seedance models to developers through BytePlus. On AdAnt you can run it without your own provider key, from the adant-cli media command or the AdAnt plugin for Claude, paying in AdAnt credits. A 5-second 720p clip costs about 7 credits.

Use Seedance 2.5 for

Seedance 2.5 compared by use case

Try another model

Compare every model

Want the whole ad, not one clip?

AdAnt's agents research your market, write the script, pick the right model for each shot and make the video in one chat.

Start with AdAnt