Film, television, and previs
Create story concepts, multi-character shorts, action transfer, relighting, style studies, extensions, and 3D white-model previs.
Available on Open Sora · Official ByteDance model
Official model guide and online generator
Seedance 2.5 AI Video Generator
ByteDance has officially introduced Seedance 2.5 with 4–30-second generation, up to 50 multimodal references, precise video editing and extension, and native audiovisual creation in more than 10 languages. It is available in the Open Sora video generator now. Use the Seedance 2.5 AI video generator directly below to create from a prompt, image, frame pair, or supported references.
Generate with Seedance 2.5 online
The complete Open Sora video workspace is embedded here and starts with Seedance 2.5 selected. You can still compare variants or switch models without losing the rest of the workflow.
Loading the Seedance 2.5 AI video generator…
Generation requires an account and uses credits based on the selected model, variant, duration, resolution, and other settings. Failed tasks are refunded automatically.
What is Seedance 2.5 and when should you use it?
Understand where Seedance 2.5 fits before spending time on references, prompts, and final-resolution generations.
Seedance 2.5 is ByteDance's latest professional multimodal video model. Its official guides document any integer duration from 4 to 30 seconds, up to 30 images plus 10 videos and 10 audio clips, audio-only references, precise editing, forward and backward extension, and stronger instruction following. Seedance 2.5 is selectable in the Open Sora generator on this page.
For film and television teams, advertising and ecommerce studios, educators, multilingual marketers, industrial teams, and enterprise creators building reference-rich video workflows, the practical advantage is not a single headline benchmark. It is the way Seedance 2.5 combines Text, image, first and last frames, multimodal reference, audio-only reference, video editing, and extension with Up to 30 images, 10 videos, and 10 audio clips (50 assets total). That combination determines whether the model can preserve a prepared visual direction or needs to invent most of the scene from language alone.
On this page, research and production live in one flow. Read the official specifications, study the source material, use the prompt framework, and then work in the embedded generator. The selected model defaults to Seedance 2.5, while the rest of Open Sora's upload, progress, history, reuse, and download experience stays available.
Seedance 2.5 specifications and supported formats
Use these official capabilities to plan the input, format, duration, and production tier before you generate.
Developer
ByteDance
Model release
Available on Open Sora
Generation modes
Text, image, first and last frames, multimodal reference, audio-only reference, video editing, and extension
Inputs
Up to 30 images, 10 videos, and 10 audio clips (50 assets total)
Duration
Any whole number of seconds from 4 to 30
Resolution
480p and 720p
Aspect ratios
21:9, 16:9, 4:3, 1:1, 3:4, 9:16, and Smart Aspect Ratio
Audio
Native audiovisual generation in more than 10 languages, plus audio-only references
Frame rate
24 FPS
Output formats
MP4
Best use cases for Seedance 2.5
Start with work where the model's strongest controls create a practical advantage rather than choosing only by maximum resolution.
Advertising and ecommerce
Produce 30-second TVCs, 3D animated ads, viral-content variations, and reusable campaigns across batches of product SKUs.
Knowledge and education
Turn concepts, source material, and narration into multilingual explainers with coordinated visuals, speech, timing, and camera direction.
Industrial and enterprise video
Build product demonstrations, operating instructions, training material, and scenario videos while preserving approved assets and structure.
Official Seedance 2.5 example and source material
This official media was downloaded from the model developer's launch or product material and optimized for fast playback on this page.
Official Seedance 2.5 preview
Official source material: Official 30-second Seedance 2.5 preview downloaded from the Volcano Ark model page and optimized for web playback.
How to create video with Seedance 2.5
Move from a creative idea to a configured Seedance 2.5 generation without leaving this model page.
1
Describe the shot
Write the subject, action, environment, camera, style, timing, and sound. If you have reference media, upload it and explain the role of each asset.
2
Configure Seedance 2.5
Keep Seedance 2.5 selected, choose the appropriate variant, mode, duration, aspect ratio, resolution, and audio settings, then check the displayed credit cost.
3
Generate, review, and reuse
Start the task, follow progress in the result panel, review the completed video, reuse its settings for another take, or download the finished file.
Why creators choose the Seedance 2.5 AI video generator
The defining capabilities that shape how Seedance 2.5 handles direction, references, motion, sound, and delivery.
4 to 30 seconds in one take
Choose any whole number of seconds from 4 to 30 — long enough to hold a full narrative arc, a product demo, or a multi-beat sequence in a single generation.
Up to 50 multimodal references
Combine up to 30 images, 10 reference videos, and 10 audio clips in one generation—50 reference assets in total.
Audio-only references and 10+ languages
Use audio alone to guide rhythm, beat timing, lip sync, voices, and effects, with native generation across more than 10 languages.
Precise video editing
Add, delete, replace, or redraw selected elements while preserving the wider composition, protagonist, timing, and camera movement.
High-fidelity video extension
Extend footage forward or backward, complete transitions, and continue the logic between clips with smoother audiovisual continuity.
More production-ready output
Improved lighting, camera movement, performance understanding, and emotional impact help reduce the synthetic look of generated footage.
How to prompt Seedance 2.5 for more controllable video
A strong Seedance 2.5 prompt behaves like a compact production brief: it gives the model a subject, an ordered action, a camera plan, an art direction, and a soundtrack.
Reusable prompt structure
Subject + ordered action + environment + camera + lighting + visual style + timing + dialogue and sound + consistency constraints
Example Seedance 2.5 prompt
“Create a continuous 30-second product film. Use Image 1 for the exact product, Image 2 for the desert palette, Video 1 for the circular camera motion, and Audio 1 for rhythm. 0–6s: slowly push toward the package as wind moves the sand. 6–15s: circle the opening product as citrus particles appear on the beat. 15–24s: move through three environments while preserving the package geometry and label. 24–30s: finish with the exact tagline, warm backlight, and a short sonic logo.”
Curate references before adding more
For greater stability, keep the main subjects to five or fewer and start with 6–10 clear assets, ideally one clear view per subject.
Keep motion and audio references focused
Use clear 5–10-second video or audio clips where possible, and state whether each controls motion, voice, rhythm, effects, or transitions.
State what editing must preserve
For a local change, explicitly lock the camera, timing, protagonist, lighting, and unaffected areas before describing the revision.
Seedance 2.5 modes, references, and output options
Choose a tier and format based on where you are in the creative process. Draft settings are for finding the shot; premium settings are for finishing a direction that already works.
| Capability | Seedance 2.5 support |
|---|---|
| Variants | One Standard tier. Credits are charged per second of output: 10 per second at 720p, 7.5 at 480p, plus 15% when audio generation is on. |
| Generation modes | Text, image, first and last frames, multimodal reference, audio-only reference, video editing, and extension |
| Inputs | Up to 30 images, 10 videos, and 10 audio clips (50 assets total) |
| Reference control | 30 images + 10 videos + 10 audio clips; up to 50 assets total |
| Duration | Any whole number of seconds from 4 to 30 |
| Resolution | 480p and 720p |
| Aspect ratios | 21:9, 16:9, 4:3, 1:1, 3:4, 9:16, and Smart Aspect Ratio |
| Audio | Native audiovisual generation in more than 10 languages, plus audio-only references |
| Frame rate | 24 FPS |
| Output formats | MP4 |
Any whole number of seconds from 4 to 30480p and 720pNative audiovisual generation in more than 10 languages, plus audio-only references
Plan around the limits of generative video
AI video is most reliable when the prompt gives each shot one readable visual idea. Review important details before publishing and treat the first generation as a directed take that can be refined.
Complex multi-character interaction, fast occlusion, readable text, logos, hands, and exact object counts can still vary between takes. Use clear references, simplify crowded action, and inspect continuity frame by frame.
Higher resolution does not replace art direction. Lock the story beat, composition, movement, and sound at an economical setting first; then move the strongest direction to the premium variant or resolution.
Seedance 2.5 AI video generator FAQ
What is Seedance 2.5?
Seedance 2.5 is ByteDance's latest professional multimodal video model. Official documentation covers 4–30-second generation, up to 50 image, video, and audio references, precise editing and extension, and native audiovisual creation in more than 10 languages.
Is Seedance 2.5 available in Open Sora now?
Yes. Seedance 2.5 is live in Open Sora and is selected by default in the generator on this page. You can also pick it from the model list anywhere else in the workspace.
Which model does the generator on this page use?
Seedance 2.5. The embedded workspace starts with Seedance 2.5 Standard at 720p with audio on. You can switch to any other model from the picker without losing the rest of the workflow.
What duration and resolution does Seedance 2.5 support?
Any whole number of seconds from 4 to 30, at 480p or 720p and 24 FPS. Higher resolutions are not available for this model.
How many references can Seedance 2.5 use?
It supports up to 30 images, 10 reference videos, and 10 audio clips—50 assets total. Each video or audio clip may be 2–30 seconds, with no more than 30 seconds of video references and 30 seconds of audio references in total.
Does Seedance 2.5 support local video editing?
Yes. Official guides describe adding, deleting, changing, or redrawing specified elements while preserving the broader composition, protagonist, timing, and camera movement. They also document forward and backward extension and transition completion.
Does Seedance 2.5 generate audio?
Yes. Seedance 2.5 creates coordinated video and audio in more than 10 languages. It also accepts audio-only references to guide rhythm, beat timing, lip sync, voices, and effects.
Which aspect ratios and output formats are supported?
Official specifications list 21:9, 16:9, 4:3, 1:1, 3:4, 9:16, and Smart Aspect Ratio, with MP4 and MOV output. First-frame, first-and-last-frame, extension, and editing workflows inherit the input media ratio through Smart Aspect Ratio.
Official Seedance 2.5 sources
Model capabilities and media on this page were researched from the developer's official product pages, announcements, and documentation.
Create your next video with Seedance 2.5
Open the complete Seedance 2.5 AI video generator above, add your prompt or references, and turn the next shot on your list into a finished video.
Start generatingBrowse all models