Seedance 2.5 AI Video Generator
Create 4–30 second videos with text, first and last frames, multimodal references, native audio, web search, and MP4 or MOV output.
Seedance 2.5 is ByteDance Seed's production-focused video model for longer continuous stories and precise multimodal control. Raphael supports text-to-video, image-to-video with first and last frames, and reference generation with up to 30 images, 10 videos, and 10 audio clips, with no more than 50 reference assets in one request. Generate at 480p or 720p, choose any whole-second duration from 4 to 30 seconds, create synchronized audio, enable web search for text prompts, and export MP4 or MOV.
What Makes Seedance 2.5 Stand Out
Create longer stories with consistent characters, production-ready compositing, and nuanced motion and emotion.
A young girl with two hair buns pilots a pink-and-white airplane through golden clouds as dragons and whales glide beside her. She then rides a surfboard across the sea, meets a small red dragon, and finally dives into a glowing underwater world. Keep the girl, her clothing, and her fantasy companions consistent through fluid transitions.
Seedance 2.5 Key Features
Longer continuous stories, stable characters and scenes, subtle emotional expression, and natural complex motion.
“A young girl with two hair buns pilots a pink-and-white airplane through golden clouds as dragons and whales glide beside her. She then rides a surfboard across the sea, meets a small red dragon, and finally dives into a glowing underwater world. Keep the girl, her clothing, and her fantasy companions consistent through fluid transitions.”
Up to 30 Seconds of Continuous Storytelling
Create videos up to 30 seconds while preserving identity, wardrobe, and story continuity across changing scenes.
“A woman in a deep blue dress slowly walks down ancient moss-covered stone steps surrounded by glowing flowers, mist, stone arches, and warm sunlight. The camera follows smoothly and moves closer while her face, clothing, and lighting remain consistent.”
Stable Character and Scene Consistency
Keep faces, clothing, lighting, and environment detail stable as the subject moves and the camera advances through a fantasy scene.
“Against a black background, a short-haired woman tries to hold back tears. Keep a frontal close-up that captures natural skin, reddened eyes, lightly trembling lips, and a tear gradually running down her cheek with restrained movement.”
Subtle Emotion and Micro-Expressions
Render breathing, eye and lip movement, and gradually forming tears for restrained, natural emotional close-ups.
“A dancer in a red dress performs controlled arm movements at the center of a dark stage. The camera pushes in from a frontal upper-body view as flames rise behind her, then cuts to a close-up of her steps and flowing skirt while keeping the motion and rhythm continuous.”
Complex Motion and Natural Camera Work
Keep the dancer's gestures, body movement, flowing skirt, and fire rhythm coherent while connecting an upper-body push-in with a close-up of her steps.
Explore other models
Switch between top video models in one generator.
MiniMax H3
2K video with native stereo audio and multimodal references.
Seedance 2.0
Smooth motion and natural choreography.
Seedance 2.0 Mini
Lower-cost drafts with rich reference control.
Seedance 1.5 Pro
Cinematic 1080p video with synchronized native audio.
Seedance 1.0 Pro Fast
Fast, lower-cost text and image-to-video up to 1080p.
Veo 3.1
Native audio, 4K output, cinematic realism.
Compare Seedance Video Models
Choose between fast drafting, synchronized audio, lower-cost references, and Seedance 2.5's longer production workflow.
| Feature | Seedance 1.0 Pro Fast | Seedance 1.5 Pro | Seedance 2.0 Mini | Seedance 2.0 | Seedance 2.5 |
|---|---|---|---|---|---|
| Resolution | 480p / 720p / 1080p | 480p / 720p / 1080p | 480p / 720p | 480p / 720p / 1080p | 480p / 720p |
| Duration | 2–12 seconds | 4–12 seconds | 4–15 seconds | 4–15 seconds | 4–30 seconds |
| Reference Inputs | Text or first frame | Text or up to 2 images | Image / Video / Audio | Image / Video / Audio | Up to 50 mixed assets |
| Native Audio | No | Yes | Yes | Yes | Yes, on by default |
| Delivery | Standard output | Standard output | Standard output | Standard output | MP4 / MOV, web search |
| Best For | Fast low-cost drafts | Audiovisual shots | Low-cost reference tests | Flagship short-form work | Longer production workflows |
Use Seedance 2.5 for longer coherent shots and advanced delivery, or choose an earlier model when speed or lower cost matters more.
Choose Your Plan
Free images include a watermark. Upgrade for clean outputs, faster generation, and commercial use.
Pro
-50%- Advanced models after upgrading
MiniMax H3
Nano Banana
GPT
SeedreamVeo
Seedance 2.5
Kling - 2,000 credits per month
- Up to 2,000 fast images
- Up to 250 basic videos
- Unlimited Seedream 3.5 generations
- Priority queue
- No watermarks
- Batch AI image upscaling
Ultimate
-50%- Advanced models after upgradingMiniMax H3Nano BananaGPTSeedreamVeoSeedance 2.5Kling
- 5,000 credits per month
- Up to 5,000 fast images
- Up to 625 basic videos
- Unlimited Seedream 3.5 generations
- Highest priority queue
- No watermarks
- Batch AI image upscaling
- Full privacy
Max
-50%- Advanced models after upgradingMiniMax H3Nano BananaGPTSeedreamVeoSeedance 2.5Kling
- 10,000 credits per month
- Up to 10,000 fast images
- Up to 1,250 basic videos
- Unlimited Seedream 3.5 generations
- Highest priority queue
- No watermarks
- Batch AI image upscaling
- Full privacy
Seedance 2.5 Frequently Asked Questions
Learn about Seedance 2.5 generation methods, duration, multimodal references, sound, and output formats.
What generation methods does Seedance 2.5 support?
It supports text-to-video, image-to-video with first and last frames, and multimodal reference generation using images, video, and audio.
How long can a Seedance 2.5 video be?
It can create videos from 4 to 30 seconds, covering short shots, continuous transitions, and more complete story beats.
How many reference files can I use?
Reference mode supports up to 30 images, 10 videos, and 10 audio clips, with no more than 50 assets in one request. Reference video and audio can each total up to 30 seconds.
Does Seedance 2.5 generate audio?
Yes. It can create dialogue, ambience, action sounds, and musical rhythm synchronized with the visuals.
What does web search add?
For text-to-video, web search can add real-world information and background details related to the prompt.
Can I export MOV?
Yes. Choose MP4 or MOV for use across different players and editing workflows.
What can multimodal references control?
Images can guide identity and visual style, video can guide motion and camera work, and audio can guide dialogue, rhythm, and atmosphere.
What content is Seedance 2.5 suited for?
It works well for continuous stories, consistent characters, emotional performances, complex action, advertising, and cinematic shots with synchronized sound.
