Ready to create some AI magic? Describe your scene, pick your settings, and watch as AI brings your vision to life in stunning video.
Seedance 2.5 is ByteDance's longest-form video model on Zyka. Generate a single continuous shot of up to 30 seconds at 480p or 720p, with native audio, up to 30 reference images, 10 reference videos, and 10 reference audio tracks.
Seedance 2.5 renders up to 30 seconds in one shot, so the prompt can carry a full beat — setup, movement, and resolution — instead of a single moment.
Upload up to 30 images, 10 videos, and 10 audio tracks, then call them in the prompt as @Image1, @Video1, @Audio1 to steer identity, motion, and sound.
Choose any duration from 4 to 30 seconds at 480p or 720p. Native audio is generated alongside the visual track in the same pass.
Seedance 2.5 is ByteDance's long-form video model on Zyka, available via fal.ai. Where the Seedance 2.0 tiers cap out at 15 seconds, 2.5 generates up to 30 seconds natively in a single pass — no stitching, no continuation calls, and no drift between segments.
It also carries the widest reference budget in the Seedance family: 30 reference images, 10 reference videos, and 10 reference audio tracks, all addressable from the prompt with @-mentions. That makes it a strong fit for scenes that need consistent characters, props, or motion style across a longer runtime.
Output is capped at 480p and 720p. When you need 1080p or 4K, use standard Seedance 2.0 — 2.5 trades peak resolution for length and reference depth.
Seedance 2.5 is ByteDance's long-form video model — up to 30 seconds of continuous cinematic video with native audio, generated in a single pass from text, images, video, or audio references.
Up to 30 seconds, selectable in one-second steps from 4 seconds upward. It is generated as one continuous shot rather than stitched segments.
No. Seedance 2.5 outputs 480p or 720p. For 1080p or 4K, use standard Seedance 2.0.
Up to 30 images, 10 videos, and 10 audio files. Mention them in the prompt as @Image1, @Video1, and @Audio1 to control which reference drives which part of the scene.
Use 2.5 when runtime and reference depth matter — longer scenes, consistent characters, richer audio direction. Use 2.0 when you need maximum resolution, including 1080p and 4K.