All models
Seedance 2.5 · by ByteDance
Seedance 2.5 AI Video Generator
Seedance 2.5 is ByteDance's newest video flagship: top quality and motion, native synced audio and dialogue, every aspect ratio from 21:9 cinema to 9:16 vertical, and single clips up to 30 seconds — twice what Seedance 2.0 renders in one pass. Generate with it online on OpenGradient — no app, no API keys, free to start.
Free credits to start · no card · lands with Seedance 2.5 preselected
Real clips generated on OpenGradient
Capabilities
What you can make with Seedance 2.5
Half-minute scenes with sound, from a single prompt — no stitching, no separate audio pass.
30 seconds in one clip
Durations from 5 to 30 seconds — room for a full ad spot, a scene with a beginning and an end, or a multi-shot sequence rendered in one pass instead of assembled from fragments.
Audio born with the video
Dialogue with natural lip-sync, ambience and effects are generated alongside the frames — the same synced-audio pipeline as Seedance 2.0, over twice the runtime.
Production-brief prompts
Prompts run to 10,000 characters, five times the 2.0 limit, so timestamped shot lists with camera, lighting and audio direction fit in one prompt instead of being cut down to size.
Up to 30 reference images
In image-to-video mode, stack references for characters, products and locations — over three times what the 2.0 endpoints accept — and the render follows your source image's framing.
Seedance 2.5 at a glance
Modes
Text to video + Image to video
Clip length
5s – 30s
Aspect ratios
16:9 · 9:16 · 1:1 · 4:3 · 3:4 · 21:9
Audio
Synced audio & dialogue
Moderation
Provider-filtered
Variants
6 live on OpenGradient
The Seedance lineup, live on OpenGradient
Where it fits
How people use Seedance 2.5
The pick when the brief needs length: a whole spot, a whole scene, a whole story beat in a single render.
Full-length ad spots
The 30-second slot broadcast and social have been built around, delivered as one clip with its soundtrack already in place.
Dialogue that goes somewhere
Not one line but an exchange — several beats of conversation with believable lip-sync, inside a single generation.
Multi-shot sequences
A timestamped shot list in the prompt becomes cuts in the render, with characters and environments holding across them.
Character-consistent series
Feed a reference stack in image-to-video mode and keep the same face, product or set across every clip in a campaign.
More AI video models on OpenGradient
Questions, answered
Frequently asked
- What is Seedance 2.5?
- Seedance 2.5 is the newest generation of ByteDance's Seedance video model. It turns text prompts or images into clips with synced audio — dialogue, ambience and effects generated together with the picture — and leads OpenGradient's video lineup on quality, motion and clip length.
- What's new compared to Seedance 2.0?
- Three things you'll feel immediately: single clips up to 30 seconds instead of 15, prompts up to 10,000 characters instead of 2,000 (long production briefs are its intended style), and up to 30 reference images per request in image-to-video instead of 9. Synced audio and the six aspect ratios carry over from 2.0.
- How do I use Seedance 2.5 on OpenGradient?
- Open OpenGradient's Video Studio from the button on this page — it lands with Seedance 2.5 preselected. Type a prompt (or upload an image in image-to-video mode), pick an aspect ratio and duration, and generate. The studio is visible without an account; generating takes a free sign-up.
- Can Seedance 2.5 animate an image?
- Yes — image-to-video is in the same picker, with clips up to 30 seconds and reference images for consistency. There's no aspect-ratio picker in that mode: the output follows the framing of the image you upload.
- What does a clip cost?
- Per-clip prices depend on the model and duration and are shown before you generate. 1,000 credits = $1; packs start at $10; you get free credits to start — no card needed.
- Should I use Seedance 2.5 or 2.0?
- Use 2.5 when length, a long brief or a big reference stack matters — it's the strongest model in the lineup. Seedance 2.0 remains the studio default and a cheaper render for clips of 15 seconds or less, and Seedance 2.0 Mini is the draft tier at roughly half its price. All three are one click apart in the same picker.
- How does Seedance 2.5 compare to Veo, Kling and Wan?
- Seedance 2.5 is the best all-rounder here and the longest single clip of any model on OpenGradient. Veo 3.1 pushes resolution to 4K, Kling v3 Pro specializes in motion fidelity, and the Wan line runs unrestricted. They share one model picker, so you can audition the same prompt across models for cents.
Render 30 seconds on Seedance 2.5
Free to start, no card, no API keys. Type a brief and hear your first clip in minutes.