
Seedance 2.5 just changed AI video generation: 30-second generations, 50 multimodal references, and perfectly synced audio. all in one single pass.

A full walkthrough of a space in a single
continuous shot.

Replace an object or polish a face without re-rendering the clip.

30 images, 10 videos, 10 audio. Feed it your moodboard.
Drive the film with 50 references
Drop in faces, products, wardrobe, locations, style frames – up to 50 multimodal inputs. Direction set in a few frames, applied to all thirty seconds
Try Seedance 2.5Edit exactly what you want
Swap the background without touching the subject. Same actor, same motion, same light – Seedance 2.5 rebuilds the scene.
What is Seedance 2.5?
Seedance 2.5 by ByteDance is an AI video generator that creates native 30-second clips from text, images, video, and audio references, with synchronized audio generated in the same pass. It is available now inside Lupa AI.
How long are Seedance 2.5 videos?
Each clip runs from 4 to 30 seconds. Set the exact duration manually or use Smart Duration to let the model choose a natural length for your scene. Longer clips use more tokens.
Can Seedance 2.5 generate audio?
Yes. Dialogue, music, and sound effects are co-generated with the visuals, so sound is always in sync. No separate audio pass or manual alignment needed.
What reference inputs does Seedance 2.5 support?
Up to 50 multimodal references per generation: 30 images, 10 video clips, and 10 audio files. Assign each reference a role, like subject, style, camera, motion, dialogue, music, or SFX, right from the prompt.
What resolution does Seedance 2.5 support?
480p and 720p at 24 FPS at launch, with 1080p expected soon. Need more detail today? Run your shots through the Lupa Video Upscaler to push them even further.




















