Seedance 2.5 from ByteDance
ByteDance's flagship video model. Full 30 second scenes with audio in a single pass.
What you can make with Seedance 2.5
Edit and extend real footage
Drop in a clip as a reference and describe the change. Seedance 2.5 can swap objects, restyle a scene, or continue a shot with the same characters and setting.
Direct a scene with dialogue
Put spoken lines in quotes and the model performs them. Lip movement, voice, sound effects, and music are generated together with the picture.
A much bigger reference set
Load up to 30 reference images, 10 video clips, and 10 audio tracks into one generation and call them out in the prompt by name, so a whole cast, product line, or location survives the cut.

Inside Seedance 2.5
720P resolution output
Render crisp, production-ready footage at up to 720P.
Flexible aspect ratios
Compose in 6 aspect ratios, from square to widescreen and vertical.
Clips up to 30 seconds
Generate variable-length shots without stitching multiple runs.
Frame and reference control
Steer motion with start and end frames and reference images for precise, repeatable results.
Video reference input
Drive style, motion, and pacing from an existing video clip, not just a still or a prompt.
Audio-driven generation
Sync lip movement, rhythm, and timing to a reference audio track for talking and performance shots.
Native synced audio
Generate video with synchronized sound baked in — no separate scoring or dubbing pass.
Production-ready output
Seedance 2.5 renders clean, high-fidelity footage built for real campaigns, products, and client work.
Create with Seedance 2.5
- 1
Select Seedance 2.5
Choose Seedance 2.5 from the model library to start generating cinematic videos.
- 2
Enter prompt and add references
- 3
Generate and download
Frequently asked questions
What is Seedance 2.5?
Seedance 2.5 is an AI video generation model available on Morphix. ByteDance's flagship video model. Full 30 second scenes with audio in a single pass.
How is Seedance 2.5 different from Seedance 2?
Seedance 2.5 generates up to 30 seconds in a single pass, double the 15 second cap of Seedance 2, and takes a far larger reference set: 30 images, 10 video clips, and 10 audio tracks versus 9, 3, and 3. It also adds video editing and extension, so it can modify or continue footage you upload. Seedance 2 still renders at higher resolutions, up to 4K, while Seedance 2.5 tops out at 720p. Both run in the same Morphix workspace, so you can switch between them per project.
Can Seedance 2.5 edit or extend an existing video?
Yes. Upload the clip as a reference video and describe what should change or what happens next. The model can replace objects, change the background or style, and continue the scene with consistent characters. Reference clips can total up to 30 seconds per generation.
How do I use Seedance 2.5 on Morphix?
Open the video studio on Morphix, select Seedance 2.5, write a prompt or upload a reference frame, and generate. Every model shares one subscription and credit pool, so there are no API keys or separate sign-ups.
What resolution and duration does Seedance 2.5 support?
On Morphix, Seedance 2.5 supports 480p and 720p output at durations from 4 to 30 seconds per generation.
Does Seedance 2.5 support reference-guided generation?
Yes. Seedance 2.5 supports reference-guided workflows on Morphix, so you can use images to steer composition, subject identity, and style.
Can I use Seedance 2.5 videos commercially?
Yes. On Morphix paid plans, videos you generate with Seedance 2.5 include commercial licensing for campaigns, products, and client deliverables.
Who should use Seedance 2.5?
Seedance 2.5 is useful for designers, marketers, creative teams, and independent creators who need polished AI-generated videos for concepts, campaigns, client work, and production work.
Explore more models
Seedance 2
ByteDance's latest video model with reference images, video, and audio inputs.

Seedance 2 Mini
ByteDance's lightweight Seedance video model, tuned for fast, low cost generation at scale.
Veo 3.1
Google's latest Veo model with synced audio and strong prompt adherence.
LTX-2
Lightricks' audio-video model that generates sound and picture together, so dialogue and motion stay in sync.
Wan 2.6
Alibaba's open-source video model — strong quality, friendly price.