Seedance 2.5 AI Video Generator
Turn text and up to four reference images into video with native audio, up to 30 seconds at 720p
Model
Usually 2-8 min
Flambé in a busy kitchen5 s · 16:9 · 720p · with sound
Tips for Seedance 2.5
Tag each reference in the prompt, e.g. “the headphones from @Image1 in the alley from @Image2”, then add the camera move and the sound.
Showcase
Seedance 2.5 in Action
See what Seedance 2.5 can create with these examples
Flambé in a busy kitchen
A chef flambés a pan in a busy restaurant kitchen, a burst of flame lights his face, the camera circles slowly around the stove. Sizzling, clanging pans and kitchen chatter.
Two references, one shot
A young woman wearing the sage green headphones from @Image1 walks through the rainy neon alley from @Image2 at night, handheld follow shot from behind, then she turns to the camera and smiles. Rain, footsteps and a faint city hum.
More video models
- Veo 3.1Generate 1080p videos with synchronized audio, music, and sound effects from text or images
- MiniMax H3Turn a prompt, opening and ending frames, or a complete set of image, video, and audio references into a coherent video with native stereo sound. Generate 4-15 second clips in 768P or 2K.
- Grok Imagine 1.5Turn a prompt or a single image into a dynamic AI video
- Hailuo 2.3Create professional videos with NCR technology and director-level camera control
Common Questions & Answers
Find out all the essential details about Seedance 2.5.
Need additional help? Contact us at [email protected]
Seedance 2.5 is ByteDance's newest video model. It turns a text prompt and up to four reference images into a single clip with native, synchronized audio, keeping characters and scenes consistent throughout.
Seedance 2.5 handles several reference images at once, supports clips up to 30 seconds, and follows longer, more detailed prompts. Seedance 2.0 tops out at 15 seconds with a single reference image.
Upload up to four images and point to them in your prompt as @Image1 through @Image4 — for example, @Image1 for the character and @Image2 for the setting. The model keeps them consistent across the shot.
Yes. It produces native audio synchronized with the video, including ambient sound and effects. You can turn audio off before generating.
Any duration from 4 to 30 seconds, at 480p or 720p. You can also pick an aspect ratio, or let Adaptive follow your reference images.
Yes. Reference images are optional — describe the scene, the camera movement, and the mood, and the model builds the shot from your prompt.
Credits are charged per second and depend on the resolution. The exact cost is shown above the Generate button before you submit.
In most cases, videos you create here are suitable for commercial use. Please review our Terms of Service for full licensing details.