Seedance 2.0 API
AI VideoSeedance 2.0 is the previous Seedance generation on OmniaKey, published under the provider ID dreamina-seedance-2-0-260128: one id covers text to video, a first frame and reference-driven shots of up to 15 seconds, always with an audio track.
- Per second of output$0.077–$0.407
- OperationsText to video
- Image to video
- Reference to video
- View tool pricing and billing rules
Your result will appear here once you run the model.
01
Where it fits
Seedance 2.0 is the previous Seedance generation on OmniaKey, published under the provider ID dreamina-seedance-2-0-260128: one id covers text to video, a first frame and reference-driven shots of up to 15 seconds, always with an audio track.
Use Seedance 2.0 when you want Seedance quality at roughly half the 2.5 rate, from $0.08 to $0.41 per second between 480P and 1080P; a 4-second 480P clip returned in about a minute in our record.
02
What it does
Text to video
dreamina-seedance-2-0-260128
Generate a short video clip from a text prompt.
Image to video
dreamina-seedance-2-0-260128
Animate a still image into a short video clip.
Reference to video
dreamina-seedance-2-0-260128
Guide a clip with reference media to keep a subject, style or motion.
03
In depth
Seedance 2.0 is the previous Seedance generation on OmniaKey, sold under the id dreamina-seedance-2-0-260128. It handles the same three job types as 2.5 — a prompt alone, a supplied first frame, and reference images — but every clip stops at 15 seconds and always ships with an audio track.
The appeal is the rate card: at 480P, 720P and 1080P this id costs less per output second than Seedance 2.5 does at the same three resolutions. It keeps 1080P, which the Fast and Mini tiers drop, so it is the middle option for anyone who needs a large frame but not a 20-second take.
We put three 5-second 720p prompts through it on 26 October 2026 alongside HappyHorse 1.1 and judged Seedance 2.0 the winner on all three. The observations below list those results, the render times, and a 4-second 480P smoke run from September.
Numbered-shot dialogue scenes
ByteDance's guide says 2.0 reads shot numbers rather than timestamps, so label beats Shot 1, Shot 2 and keep each under a few seconds. Give the camera move and the line of dialogue inside the shot block, then let the model write the voice track.
Product close-ups at 1080P
When the deliverable is a large frame but the cut is short, this id gives 1080P at the gentlest rate in the Seedance family. Our skate, noodle and perfume prompts at 720p all came back with readable signage and clean product edges.
Reference-driven continuity
Attach reference images through input.media and describe only the new action; the September smoke run passed an inlined reference image at 480p without trouble. Keep request bodies small, since the whole payload has to stay under 1 MiB including Base64.
04
Hands-on notes
In our same-prompt video test on 26 October 2026, Seedance 2.0 beat HappyHorse 1.1 on all three 5-second 720p prompts we sent.
Read the testThe three Seedance 2.0 clips rendered at a median 184.7 seconds against 97.6 seconds for HappyHorse 1.1 on the same prompts.
Read the testOur 14 September 2026 smoke run rendered a 4-second 480p Seedance 2.0 text-to-video job as an 864x496 H.264 MP4 of 4.04 seconds.
In the same smoke run, Seedance 2.0 reference-to-video took an inlined reference image at 480p and returned a 4.04-second 864x496 MP4.
A Seedance 2.0 reference job we sent with only an audio reference and no image or video failed instead of rendering.
05
Limits to plan around
- Fifteen seconds is the ceiling, and duration -1 will not buy more. A 20- or 30-second take belongs on Seedance 2.5, which is the only id in the family that accepts it.
- ByteDance's prompt guide reports that 2.0 ignores timestamp ranges and follows shot numbers, so a timeline written for 2.5 renders differently here until the beats are relabelled.
- On identical prompts this id finished well behind HappyHorse 1.1 in wall-clock terms, so batch jobs overnight rather than waiting on a single clip inside an editing session.
- A reference request needs at least one image or video among its inputs; audio on its own is refused before a job starts.
06
Compare nearby models
| Model | Best for | Operations | Listed price |
|---|---|---|---|
| Seedance 2.0This page dreamina-seedance-2-0-260128 | The 1080P middle tier when fifteen seconds is enough. | Text to video · Image to video · Reference to video | $0.077–$0.407 per second |
| Seedance 2.5 dreamina-seedance-2-5-260628 | Longer takes, reference work and a higher per-second rate at each shared resolution. | Text to video · Image to video · Reference to video | $0.1133–$0.6259 per second |
| HappyHorse 1.1 happyhorse-1.1 | Three ready-made variants with 480P, stronger prompt handling on some shots. | Text to video · Image to video · Reference to video | $0.042–$0.108 per second |
07
Sources and verification
Guide updated 2026-10-27.
08
Call it
Media models run asynchronously. Create a task, poll it until it finishes, then download the signed asset URL.
# 1. Create the job. 202 Accepted; the body carries the task id,
# the pre-charged price, and a poll_url. Reuse one Idempotency-Key
# across retries so a timeout never starts a second paid job.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks \
-H "Authorization: Bearer $OMNIAKEY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: dreamina-seedance-2-0-260128-001" \
-d '{
"model": "dreamina-seedance-2-0-260128",
"type": "video",
"operation": "text_to_video",
"input": {"prompt": "a paper boat on a quiet lake at sunset"},
"parameters": {"duration": 6, "resolution": "720p"}
}'
# 2. The same model id and the same poll loop serve the other operations:
# image_to_video media[] type "first_frame"
# reference_to_video media[] type "reference_image"
# Reference media travel alongside the prompt:
# {"prompt": "...", "media": [{"type": "first_frame", "url": "https://example.com/first-frame.png"}]}
# A public HTTPS URL or a base64 data URI both work.
# 3. Poll GET /v1/media/tasks/{id} until status is
# succeeded | failed | canceled. Progress carries percent and stage.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks/TASK_ID \
-H "Authorization: Bearer $OMNIAKEY_API_KEY"
# 4. Download assets[0].url before expires_at, or poll again for a
# fresh signed URL. Failed and canceled jobs are refunded.
curl -sS -o out.mp4 "$(jq -r '.assets[0].url' result.json)"09
Pricing
Video jobs are billed per second of rendered output, not per token. Failed or canceled jobs are refunded automatically.
$0.077–$0.407per second
- 480P
- $0.077per second
- 720P
- $0.165per second
- 1080P
- $0.407per second
More models from ByteDance
Related articles
HappyHorse 1.1 vs Seedance 2.0: Same-Prompt Test and Price
Same three prompts, same day: Seedance 2.0 made the better clips, HappyHorse 1.1 made them twice as fast.
Read articleSeedance 2.5 Prompt Guide: Structure, Examples and Real Clips
Six prompts, six real clips: what Seedance 2.5 followed, what it changed and how to fix it.
Read articleKling 3.0 vs Seedance 2.5
Both models are live on OmniaKey and both can serve the same brief. One is billed once per generation from a mode and duration matrix, the other per output second by resolution — and only one of them accepts a reference-driven shot.
Read articleBest AI video API models in 2026
There is no single best video model. The right route depends on whether you need a long shot, a faithful first frame, reference images, predictable cost, or a faster production queue.
Read articleFrequently asked questions
How do I call Seedance 2.0 on OmniaKey?
POST to /v1/media/tasks with the model id, type "video", and operation "text_to_video" — or "image_to_video" with a first frame in input.media[] with type "first_frame". Reuse one Idempotency-Key across retries so a timeout never starts a second paid job.
What does Seedance 2.0 cost?
Seedance 2.0 bills $0.077–$0.407 per second of rendered video. Creation pre-charges the requested duration up front from your prepaid balance, and failed or canceled jobs are refunded automatically.
Why does the API return a task id instead of the video?
Video renders take from about a minute to several minutes, so OmniaKey runs every video route through one asynchronous lifecycle. Poll GET /v1/media/tasks/{id} until the status is succeeded, then read the signed asset URL.
How does Seedance 2.0 differ from Seedance 2.5?
Three things matter: length, resolution mix and prompt syntax. 2.0 accepts 4 to 15 seconds, 2.5 goes to 30. Both price 480P, 720P and 1080P, with 2.5 charging more per second at each tier. ByteDance's guide also has 2.0 reading shot numbers where 2.5 reads timestamps.
Is Seedance 2.0 cheaper than Seedance 2.5?
Yes, at every resolution the two share. The catalog lists 2.0 below 2.5 for 480P, 720P and 1080P per output second, and both are billed on the seconds you request, taken from balance when the job is created. A failed job is refunded, so a wasted draft costs time rather than credit.
What resolutions and durations does Seedance 2.0 accept?
480P, 720P and 1080P on this id, and durations from 4 to 15 seconds. The catalog charges one rate per output second at the resolution you pick, taken from balance when the task is created and returned if the task fails, with current figures in the price table on the model page. Our smoke run asked for 4 seconds at 480p and got a 4.04-second file at 864x496, so expect the encoder to round the length up slightly.
Does Seedance 2.0 add sound to every clip?
The clips we generated all arrived with an audio track, and the model writes dialogue, effects or music from the prompt. Describe the sound you want inside the prompt, and strip the track afterwards if the edit needs a silent master.
How did Seedance 2.0 do against HappyHorse 1.1?
Better on picture and worse on waiting. On 26 October 2026 we sent the same three 720p prompts to both and preferred the Seedance clips for the skate jump, the noodle stall and the perfume orbit, while HappyHorse returned its files in less than half the median time.
