Build a 480p, 720p, or 1080p Seedance 2.5 video from text, guide frames, or a high-capacity multimodal bundle, with explicit output format and optional final-frame return.
The final task cost is rounded up to whole credits.

Example output
Seedance 2.5
Text to Video · Image to Video

Example output
Seedance 2.5
Best for
Supports
Online trial
No results to display yet.
Model examples
Explore media and text examples.
Decision fit
Decision fit
Best-fit tasks
Use it when the task looks like this
Model details
Provider
ByteDance
Category
Video
Capabilities
Text to Video · Image to Video · Reference to Video · Video to Video
Credit model
Credits depend on inputs and settings. Check the estimate before generating.
Input path
Up to 50 reference files
Prompt setup
Up to 30,000 chars
Developer access
Seedance 2.5 has callable Public API modes, but reference-media modes still need Files API upload support. Check the model reference before submitting.
Seedance 2.5 FAQ
Choose Seedance 2.5 when the workflow needs 1080p, up to 30 seconds, automatic duration, MP4 or MOV output, returned final frames, video-guided transformation, or more than nine images, three videos, or three audio files. Mini remains the lower-rate option for smaller 480p or 720p drafts up to 15 seconds.
Text accepts no uploads. Frames accepts exactly one starting image or two images for a start-and-end pair and rejects video or audio. Reference requires at least one uploaded image, video, or audio and is the only scene for a mixed multimodal bundle.
The reference scene accepts up to 30 images, 10 MP4 or MOV videos, and 10 MP3 or WAV audio files, with no more than 50 files in total. Images are limited to 30 MB each. Rivya currently limits each video to 95 MB and 2 to 30 seconds, with 30 video seconds in total; audio is limited to 15 MB and 2 to 30 seconds each, with 30 audio seconds in total.
Images and reference videos must each have width and height from 300 to 6,000 pixels and an aspect ratio from 0.4 to 2.5. Reference video must also contain 409,600 to 927,408 pixels in total and run at 24 to 60 frames per second. Prompts can contain up to 30,000 characters.
The no-video rate multiplies requested output seconds: 28 credits per second at 480p, 63 at 720p, or 158 at 1080p. Images or audio without a video do not switch the request to a with-video tier.
With a reference video, the rate is 17 credits at 480p, 38 at 720p, or 95 at 1080p, multiplied by verified input-video seconds plus reserved output seconds. Rivya preserves the 95 rate during multiplication and rounds the combined total up only once.
Automatic duration is the -1 choice. Because the final output length is not known before the run, Rivya reserves the 30-second output maximum; a video-guided request also includes verified input-video seconds. When valid actual usage at or below the reservation is reported, Rivya settles to it. Missing, invalid, or higher-than-reserved usage remains in reconciliation instead of triggering a hidden extra charge.
Choose text, frames, or multimodal reference input; set 480p, 720p, or 1080p; select a 4-to-30-second integer or automatic duration; and choose MP4 or MOV output.
No. The frames scene accepts only one or two images. Use the reference scene for image, video, or audio bundles; the two input styles are mutually exclusive.
Images may be JPEG, PNG, WebP, BMP, GIF, or TIFF. Videos must be MP4 or MOV. Audio must be MP3 or WAV. Every upload also has size, duration, geometry, and total-count checks before it can be used.
Yes. At 1080p, the no-video rate is 158 credits per output second, while a request containing video uses 95 credits across verified input-video seconds plus reserved output seconds.
Use automatic duration when the scene or source clip should guide the output length. Rivya reserves 30 output seconds in advance, so a fixed 4-to-30-second choice gives a more predictable reservation when you already know the needed length.
Use this page to validate one scene and reference hierarchy. Move to Studio when the project needs saved iterations, several source bundles, format comparisons, or a longer edit thread.
Compare alternatives
ByteDance's full Seedance 2.0 video model with explicit support for prompt-only generation, frame-driven animation, and multimodal reference generation. Rivya keeps the documented role split explicit so frame inputs and multimodal references stay mutually exclusive instead of collapsing into one ambiguous upload bucket.
Why consider it
Consider it when your next task needs Prompt, frames, or image/video/audio references.
ByteDance's faster Seedance 2.0 video model with full scene routing for prompt-only generation, frame-driven image animation, and multimodal reference video generation. Rivya keeps the documented scene split explicit so first/last-frame inputs do not collide with reference image, video, and audio roles.
Why consider it
Consider it when your next task needs Prompt, frames, or image/video/audio references.
Seedance 2.0 Mini is Rivya's lower-rate multimodal Seedance workflow for prompt-only video, first- or first-and-last-frame animation, and mixed image, video, or audio references. Its 480p and 720p tiers cover 4-to-15-second output while keeping the three input scenes mutually exclusive.
Why consider it
Consider it when your next task needs Up to 15 reference files.