Announced by ByteDance · June 23, 2026 · FORCE Conference
Seedance 2.5
One take. Thirty seconds. Native 4K.
Seedance 2.5 is ByteDance's next video generation model: a single 30-second pass with scene changes, tempo shifts, and synchronized sound — no stitching, no post. Atlas Cloud is preparing day-0 API access. Get in line before it goes public.
The Headline Feature
Thirty seconds without a single cut
Today's video models top out at a few seconds per pass, and longer stories are stitched together in post. Seedance 2.5 generates the full 30-second arc in one native pass — camera moves, scene changes, and tempo shifts included.
Establish
The opening shot sets scene, light, and mood from your prompt and references.
Scene change
The model transitions locations mid-take while characters and grade stay locked.
Tempo shift
Pacing accelerates or breathes — motion and score shift together, in sync.
Payoff
The take resolves as one continuous shot. No seams to hide, nothing to stitch.
What's New in 2.5
Built like a production pipeline, not a clip generator
Every headline capability announced for Seedance 2.5, in one place.
50 reference inputs
Mix up to 50 images, clips, and audio tracks in a single request — lock characters, products, locations, and style across the whole take. Seedance 2.0 accepts 12.
Native 4K, held for 30 seconds
Draft at 720p, iterate at 1080p, master at native 4K — and keep that fidelity across the full 30-second take, no upscaler required.
Sound and picture, one latent space
Audio is co-generated with the visuals, not layered on after. Dialogue lands on lips, foley lands on frames, the score follows the cut.
Local re-draw editing
Change one element of an existing clip — a product, a costume, the weather — while motion, camera, and lighting stay untouched. No full re-render.
Multilingual on-screen text
Titles, captions, and signage render cleanly across languages — built for localized cuts and global campaigns.
30-second consistency
Optimized spatial-temporal attention holds character appearance, lighting, and motion style across the full duration of the take.
2.0 → 2.5
The spec sheet, side by side
Announced figures for Seedance 2.5 against Seedance 2.0 as it runs on Atlas Cloud today.
Seedance 2.5 figures are ByteDance's announced specifications from the June 23, 2026 FORCE conference; independent benchmarks are not yet available. For context, Seedance 2.0 currently ranks #1 for text-to-video with audio on the Artificial Analysis Video Arena (Elo 1218, June 2026).
Sources: ByteDance / Volcano Engine announcement · Artificial Analysis Video Arena
July on Atlas Cloud
Don't wait to ship. Seedance 2.0 is 20% off all July.
The teams that get the most out of 2.5 on day one will be the ones already running 2.0 in production. Through July 31, the Seedance 2.0 and Seedance 2.0 Mini API is 20% off on Atlas Cloud — the same API shape 2.5 is expected to follow.
Discount applies automatically at billing — no code needed. Ends July 31, 2026.
Why Atlas Cloud
The platform 2.5 will land on
Atlas Cloud is where teams run ByteDance video models in production today — and where Seedance 2.5 goes live the moment access opens.
LLMs, image, video, audio, and 3D — a single key and a single invoice.
New models go live the moment they launch. Seedance 2.0 did — 2.5 will too.
Enterprise SLA, SOC II compliance, and dedicated infrastructure options.
Switch by changing a base URL — no SDK overhaul when 2.5 arrives.
Be first in line when Seedance 2.5 opens
ByteDance has 2.5 in enterprise beta now, with public rollout expected in July. Atlas Cloud will open API access the moment it's granted — and the first test cohort gets the first slots. Tell us what you're building and we'll reach out the moment it matters.
- 01Priority access when the Seedance 2.5 API opens on Atlas Cloud
- 02Launch pricing and the full spec sheet, first in your inbox
- 03Optional 2.0 → 2.5 migration review with a solutions engineer
- 04Zero commitment — one email when it matters, nothing else
Rollout timing and final capabilities are controlled by ByteDance. Joining the list secures your place in the queue — it doesn't commit you to anything, and we can't guarantee a specific date.
FAQ
Seedance 2.5, answered
Seedance 2.5 is ByteDance's next-generation AI video model, announced on June 23, 2026 at the Volcano Engine FORCE conference. Its headline capabilities: a single native pass of up to 30 seconds with scene changes and tempo shifts, up to 50 multimodal reference inputs, output from 720p to native 4K, joint audio-visual generation in one latent space, local re-draw editing, and multilingual on-screen text rendering.
The next take is thirty seconds long.
Get in line for 2.5, or start shipping on 2.0 today — 20% off through July.
