Video ● Operational Service status →

Seedance 2.5

⧉

seedance-2.5

24H STATUS Success: 100.0%

-24hnow

Model Type 2.5 Temporary availability · per clip Beta

Seedance 2.5 is the video model from ByteDance. One endpoint, parameters select the mode: text-to-video without a reference image, image-to-video with reference images (they set the subject, scene and style; they are not used as the first frame). Duration 4-30 seconds in one-second steps, 480P / 720P / 1080P, five aspect ratios (16:9 / 9:16 / 1:1 / 4:3 / 3:4), native audio available. Asynchronous: submit, then poll the job id for a re-hosted mp4 link (kept for 15 days). Billed per second; failed jobs are refunded in full.

Text-to-videoImage-to-videoUp to 30 sAsync jobs

Get API Key → API Docs

Service Highlights

4-30 seconds, selectable per second
Text-to-video and image-to-video (reference images set the subject, scene and style)
Native audio available
Billed per second, failed jobs refunded in full

Measured on real calls

Customer requests over the last 7 days

Success rate84.8%

Median render time per clip8 min

90% of jobs finish within19.9 min

Pricing

720P per second$0.158

480P per second$0.10~~-37%~~

1080P per second$0.33~~+109%~~

Shortest 4s$0.632

Longest 30s$4.740

Billed per second = rate x duration (4–30s, any integer). Per-second rates by resolution are listed above. The hold at submit equals the final settle exactly; failures are refunded in full.

View detailed pricing

Sample clips

All made with KovAPI. Click a clip to hear it.

Beneath the Walls of Troy HD original

Myth Awakens HD original

Grandma's Birthday HD original

Live Test · Playground

Try out Seedance 2.5 right here (available after login).

Input

Prompt

"Beneath the Walls of Troy" - an epic action short. @Image1 is a Greek hero in golden armour and a crimson cloak with a long spear and a lion shield; @Image2 is a Trojan champion in black armour and a blue cloak with a curved sword and a horse shield; the setting follows the walls of Troy and the sunset battlefield of @Image3. Keep each warrior's armour, helmet, cloak colour and shield emblem exactly as in their own reference - never swap them.
Opening: wind drives the sand as the two face off twenty paces apart, sharp eyes behind their helmets.
Middle: they charge at once; the spear strikes the horse shield in a burst of sparks; rapid slashes and blocks up close as the camera orbits them 360 degrees.
Climax: the setting sun sits right between them as they clash mid-air in extreme slow motion, sparks, sand and the crimson cloak hanging in the air.
Ending: both lower their weapons and salute each other.
Cinematic action-commercial look, 24fps film grain, fluid and weighty fighting; war drums, clashing metal and an epic score with the big hits on the beat. No blood or wounds, no extra characters, no on-screen text or watermarks, no deformed weapons or limbs.

0 / 11000 chars

Duration

Resolution

Aspect ratio

Reference images0 / 30

Click to upload or drag files here jpg / png / webp / gif, up to 16 MB each or paste links

With multiple references, address them in the prompt as @Image1 / @Image2, e.g. "the character from @Image1 stands in the scene from @Image2". Without tokens the model decides on its own.

AI-generated characters work fine; photos of real people may be rejected upstream (real-face support is not listed for this model). For real people use the seedance-2.0 models or seedance-2.5-30s.

Reference videos0 / 10

Click to upload or drag files here mp4 / mov, up to 50 MB and 30 s each; all reference videos together at most 30s or paste links

Can be combined with reference images: photos set the look, the video sets motion and camera. Address it as @Video1 in the prompt.

Reference audio0 / 10

Click to upload or drag files here mp3 / wav / m4a, up to 15 MB and 30 s each or paste links

Needs a reference image or video as well. Address it as @Audio1 in the prompt.

We recommend about 2–10 seconds per reference audio clip; clips longer than 10 seconds may carry a higher risk of processing failure. This is reliability guidance, not a new hard limit. Longer clips within the existing limits can still be submitted, and shorter clips do not guarantee successful generation.

Uploaded references are kept for 7 days.

Rendering usually takes 5-20 min; when the backend is busy it can exceed 60 min (no result after 120 min counts as a timeout, fully refunded). Poll the job id rather than resubmitting - each resubmission is billed separately. Billed per second, fully refunded on failure. Prompt limit 11000 characters (Chinese and English count the same).

Output

video

Results will appear here after running.

🕑 Generated images and videos are kept for 15 days before 2026-10-29 09:49:40 UTC; from that time, they are kept for 7 days and existing files older than 7 days are removed.

Content requirements

Full details →

Everything is reviewed
The prompt, the reference files and the finished clip all go through content review, picture and sound alike.
No charge if rejected
A rejected job is not charged; the credits held at submit are refunded in full.
The finished clip is reviewed again
Some jobs are rejected at submit; others only when the finished clip is reviewed, anywhere from a few minutes to over half an hour later.
Sound is reviewed too
Spoken lines, narration and lyrics in your prompt are voiced or sung, and reviewed together with the picture. If you need no voices, leave out lines and lyrics. This model always adds a soundtrack; it cannot be turned off. Uploaded reference audio is reviewed as well.
Commonly rejected
Sexual content or nudity; gore, self-harm and violent injury such as gunshot wounds; firearms, explosives and drugs; minors shown with alcohol, tobacco or in danger; politicians, celebrities and other public figures, in looks or voice (look-alikes included); copyrighted characters.
One bad shot fails the whole clip
If a single shot in a long storyboard script breaks the rules, the whole clip is rejected. Resubmitting the same input is rejected again; fix that part or replace the reference, then retry.

About Seedance 2.5

Seedance 2.5 is ByteDance's AI video model and the longest-form video model on KovAPI. One model, one set of params: omit the reference image for text-to-video, include reference images for image-to-video (they set the subject, scene and style; they are not used as the first frame). Duration is any integer 4-30 seconds, resolution 480P / 720P / 1080P, five aspect ratios (16:9 / 9:16 / 1:1 / 4:3 / 3:4), with an optional generated soundtrack. Async job: POST /v1/videos/generations returns HTTP 202 and a job id immediately; poll GET /v1/videos/jobs/{id} until status=succeeded and read data[0].url. That link is an mp4 re-hosted by us — it is kept for 15 days (far longer than the upstream CDN link), so download it or copy it to your own storage before then.

Use cases

Long takes and full narratives

Up to 30 seconds — the only model here that can carry a complete story in one shot, ideal for concept films and product narratives.

Image-to-video that keeps the subject

Reference images pin down the subject, wardrobe and scene, so product shots and character sheets come alive; name them @Image1 / @Image2 in the prompt and the identity carries through.

Precise camera control

Push in, pull out, pan and tilt written in the prompt are followed closely; pick this when the camera language matters.

Vertical and ultra-wide social

9:16 for short-video platforms, 21:9 for banners and trailer-style pieces.

How to choose

Choose 2.5 when you need more than 15 seconds or explicit camera control. Within 15 seconds and without strict camera requirements, seedance-2.0-fast is faster and cheaper; for resolution and native audio look at minimax-h3. 480P and 720P are priced separately (prices are on this page): use 480P for drafts and smaller files, 720P for the final cut.

FAQ

How does one model make different video types?

Params decide: omit image for text-to-video, include it for image-to-video; duration sets length, resolution sets sharpness, size sets the aspect ratio, audio toggles the soundtrack.

480P or 720P?

Each resolution has its own per-second price, listed on this page. 480P suits drafts, previews and smaller files; use 720P for the final cut. Either way you pay only for clips you receive, and a failed job is refunded in full.

How does image-to-video work?

Pass images (an array of reference images; a single one can go in image), as public URLs, data: URIs or base64. They reach the model as references for the characters, objects, scene or style, and you address them in order as @Image1, @Image2 in the prompt; the model has no first/last-frame control, so a reference image is not guaranteed to be the opening frame. The same request can also carry reference videos in video_references (@Video1, for motion and camera) and reference audio in audio_reference (@Audio1).

How is it billed, and am I charged on failure?

Billed per second: the hold placed at submit is computed from the duration you asked for, and the settle is the exact same number. A failed or timed-out render refunds the whole hold, so you never pay for a clip you did not receive.