Seedance 2.5: Your Imagination, Frame by Frame
Seedance 2.5 turns 50 multimodal inputs into 30s native single-take clips — with localized editing and 3D white-model rendering. No more 15s stitching, no jump cuts. Join the waitlist to be first in line.
2847 creators already on board
What Seedance 2.5 Upgrades
Seedance 2.5: 30s Full Video Output
Seedance 2.5 outputs a full 30 seconds in one native take — no more 15s stitching, no jump cuts.
- One take covers scene cuts, spatial transitions, rhythm shifts, and a thematic resolution — narrative continuity held end to end
- Every transition stays seamless, so the piece reads as one continuous performance rather than a collage glued together in post
- One native take means a single render bill — a 30-second story often costs less than three stitched 10-second clips
Screenshots from official release video
Screenshots from official release video
Seedance 2.5: 50 Multi-Modal Inputs
Feed Seedance 2.5 up to 50 multi-modal references at once — what you put in is what you get back.
- Text, images, video, and audio freely combined — an image sets style, text sets action, audio sets rhythm
- Every input stays consistent across the full 30-second render; conflicts between references are resolved on the fly
- A storyboard image, a voiceover clip, and a mood board coexist in one shot — the intent you start with is the intent you finish with
Seedance 2.5: Localized Editing
Seedance 2.5 supports precise localized editing — fix one flaw without regenerating the entire video.
- Touch up a single frame and the rest is preserved — swap a wrong prop, fix a hand distortion, or correct a lighting slip without re-running the shot
- Edits stay isolated, so a one-frame fix never ripples into the surrounding footage
- The cost of a small fix stays small — a note on frame 240 no longer means re-rendering frames 1 through 239
Screenshots from official release video
Screenshots from official release video
Seedance 2.5: 3D White-Model Rendering
Feed Seedance 2.5 a 3D white model and it renders into a fully shaded shot — structure, proportions, and motion stay under your direction.
- A controllable path from grey model to final frame for film and ad teams, with a clean handoff into live-action compositing
- Directors iterate on framing and timing in the white model before committing to a full render
- The white model becomes a reusable asset — plan once, render many; the final render rarely surprises the team that reviewed it
Seedance 2.5 Scenarios & Use Cases
Screenshots from official release video
Brand Ads with Seedance 2.5
Feed Seedance 2.5 your brand assets and multi-shot instructions — brand elements stay locked across every frame, one shoot to multi-format output.
- One asset exports to horizontal, vertical, and square, with brand colors and logos locked across each cut
- Palette, typography, and logo placement carry over between formats without manual rework
- A single 30-second master seeds a full campaign — a 16:9 hero spot, a 9:16 story, and a 1:1 feed ad, all on brand
E-Commerce with Seedance 2.5
Keep the original shots and ad-grade look with Seedance 2.5, then swap local details — one asset spins into multiple market versions fast.
- Use the same product hero, swap backgrounds, copy, and model poses to generate versions for multiple regions
- Local market teams get tailored assets without a full reshoot, cutting both cost and turnaround time
- A holiday campaign can reuse the same hero shot with only a background and copy swap
Screenshots from official release video
Series & Shorts with Seedance 2.5
30s straight output with multi-input — control camera, blocking, and interaction via prompt while Seedance 2.5 keeps characters consistent across shots.
- Character features lock across cuts, supporting multi-episode series and short-drama output
- Wardrobe, face, and voice cues persist, so continuity holds across episodes and cuts
- A recurring lead appears episode after episode without a costume or face drift breaking the storyline
Film Pre-Vis with Seedance 2.5
Feed Seedance 2.5 a ~100k-poly white model plus material refs and it turns into fully rendered video.
- Structure, proportions, and motion stay stable through impact scenes
- The render stays fast and predictable, carrying upstream design assets and shot plans forward
- Pre-vis is cheap to iterate — a director validates a stunt beat in grey model before any expensive rendering begins
Screenshots from official release video
IP Remix with Seedance (2.0 supported)
The Volcano Engine AI Copyright Platform bridges IP holders and creators — Seedance powers AI creation templates that now drive 100,000+ creations per day.
- Partnered with Stephen Chow's Beyond Group, the licensing pipeline lets compliant remixes scale up without legal friction
- Creators get ready-made templates built on licensed footage, lowering the bar for high-quality remixes
Social Shorts with Seedance 2.5
Fast turn-around with full style control — Seedance 2.5 turns one idea into platform-ready cuts in minutes.
- Output adapts to 9:16, 1:1, and 16:9, matching each platform's duration and bitrate
- A single prompt branches into platform-specific cuts, each tuned to where it will be watched
- One render ships a TikTok vertical, a YouTube horizontal, and a square feed post
Seedance Official Showcase
Seedance 2.0 official showcase — released alongside 2.5.
Seedance 2.5 lands in early July 2026. Join the waitlist to be first in line.
Seedance Pricing
Official published pricing from model providers. Estimates based on 16:9 aspect ratio, 5s output.
Seedance 2.5 (Coming Soon) | Dreamina-Seedance-2.0 | Dreamina-Seedance-2.0-fast | Dreamina-Seedance-2.0-mini | ||
|---|---|---|---|---|---|
| 480p | Without video input | ? | $0.35/unit Token: $7.0/M | $0.28/unit Token: $5.6/M | $0.18/unit Token: $3.5/M |
| With video input (2–15s) | ? | $0.39–0.86/unit Token: $4.3/M | $0.30–0.66/unit Token: $3.3/M | $0.19–0.42/unit Token: $2.1/M | |
| 720p | Without video input | ? | $0.76/unit Token: $7.0/M | $0.60/unit Token: $5.6/M | $0.38/unit Token: $3.5/M |
| With video input (2–15s) | ? | $0.84–1.86/unit Token: $4.3/M | $0.64–1.43/unit Token: $3.3/M | $0.41–0.91/unit Token: $2.1/M | |
| 1080p | Without video input | ? | $1.87/unit Token: $7.7/M | — | — |
| With video input (2–15s) | ? | $2.06–4.57/unit Token: $4.7/M | — | — | |
| 4K | Without video input | ? | $3.89/unit Token: $4.0/M | — | — |
| With video input (2–15s) | ? | $4.20–9.33/unit Token: $2.4/M | — | — |
Seedance 2.5 FAQ
Everything you need to know before getting started.