Up to 30-second engine output
The Seedance 2.5 API accepts integer durations from 4 to 15 seconds, with 15 seconds as the generation cap.
Seedance 2.5 generation is live
The playground now submits directly to the seedance-2.5 task type with the confirmed modes, resolutions, reference limits, audio control, and pricing.
Seedance 2.5 on PiAPI
More story.More control.
Create longer, reference-rich AI video with three generation modes, 480p or 720p output, and image, video, and audio inputs.
Updated July 31, 2026
- duration range
- 4–15 sec
- default output resolution
- 720p
- text, frames, and omni reference
- 3 modes
Seedance 2.5 capabilities
A focused contract for richer video workflows
Seedance 2.5 keeps the proven Seedance request shape while defining a clear model, resolution, duration, reference, and pricing contract.
480p and 720p output
Generate at 480p or 720p, with 720p as the default. Seedance 2.5 does not expose a 1080p option.
Three generation modes
Use text_to_video, first_last_frames, or omni_reference, with the same automatic mode inference pattern.
Multimodal references
Combine up to 12 references: up to 9 images, 3 videos, and 3 audio files within the total limit.
One focused model
Use the single seedance-2.5 task type without fast, mini, or less-restriction model variants.
Audio-aware generation
Control audio generation in the request and combine image, video, and audio signals in omni-reference workflows.
Confirmed model contract
Seedance 2.5 specs and pricing
The live playground uses the dedicated model rates and the single Seedance 2.5 variant shown here.
Read the Seedance 2.5 API specificationsBilling formula
Output duration × selected rate, plus input video duration × half of the selected rate.
480p output
$0.30
per output second
720p output
$0.60
per output second · default
- Task type
- seedance-2.5
- Resolution
- 480p or 720p; 720p default; no 1080p
- Duration
- 4–15 seconds · integer values
- Generation modes
- text_to_video · first_last_frames · omni_reference
- Reference limits
- 12 total · 9 images · 3 videos · 3 audio files
- Model variants
- Single model; no fast, mini, or LR variants
- Private assets
- asset:// is not supported
- Prompt
- Up to 4,000 characters
A richer generation canvas
Combine image, video, and audio direction
Omni-reference mode accepts up to 12 total inputs, with individual caps of 9 images, 3 videos, and 3 audio files. Use the familiar Seedance prompt and aspect-ratio structure.
- Write prompts up to 4,000 characters
- Use up to 9 image references within the total limit
- Add up to 3 video and 3 audio references
- Choose text, first/last-frame, or omni-reference mode
Reference stack
One direction, multiple signals
01
Images
Up to 9 references
02
Videos
Up to 3 references
03
Audio
Up to 3 references
04
Prompt
Up to 4,000 characters
05
Modes
3 generation workflows
06
Output
480p or 720p
From playground to API
Start shaping the workflow today
Test the exact Seedance 2.5 modes in the playground, then reuse the same request contract in your application.
01Live
Try the live playground
Explore prompts, the three generation modes, reference inputs, audio, ratios, duration, and resolution.
02Connected
Generate with Seedance 2.5
Submit directly with model seedance and task_type seedance-2.5 using the confirmed input contract.
03API ready
Integrate the same request
Carry the tested mode, prompt, media URLs, resolution, duration, aspect ratio, and audio setting into your app.
Frequently asked questions
Seedance 2.5 on PiAPI
What is Seedance 2.5?
Seedance 2.5 is the next generation of the Seedance AI video model family. The confirmed PiAPI contract uses one seedance-2.5 task type with 480p and 720p output, three generation modes, multimodal references, and a duration range of 4 to 15 seconds.
Can I use the Seedance 2.5 playground on PiAPI?
Yes. The Seedance 2.5 playground is live on this page and submits generation tasks through PiAPI using the seedance-2.5 task type.
Does the playground use the dedicated Seedance 2.5 endpoint?
Yes. Playground requests use model seedance with task_type seedance-2.5 and the confirmed Seedance 2.5 input contract.
What can Seedance 2.5 create?
Seedance 2.5 supports text-to-video, first/last-frame generation, and omni-reference generation with image, video, and audio inputs. It outputs 480p or 720p video and supports an audio generation switch.
How much does Seedance 2.5 cost?
Seedance 2.5 costs $0.30 per output second at 480p and $0.60 per output second at 720p. Input video duration is billed at half the selected output rate.
Can I integrate the Seedance 2.5 API now?
Yes. Create a PiAPI task with model seedance and task_type seedance-2.5, then provide the prompt, mode, duration, resolution, aspect ratio, reference URLs, and audio setting required by your workflow.
Does Seedance 2.5 support private asset:// references?
No. Seedance 2.5 does not have a less-restriction variant, so private asset:// references are not supported in the confirmed contract.
Create with more control
Take Seedance 2.5 from idea to interface
Generate with the dedicated Seedance 2.5 task type now, then use the same request contract in your application.
Try Seedance 2.5