AI video model that generates a 30-second clip in one pass — no stitching.
Last verified · 2026-06-24 · by Moe Ameen
Seedance 2.5 is ByteDance's text- and image-to-video model, unveiled at the company's Volcano Engine FORCE conference on June 23, 2026. It is the engine behind ByteDance's creative apps, including Dreamina (the international Jimeng), the Doubao assistant, and CapCut, and it is distributed to businesses through Volcano Engine, ByteDance's cloud platform.
The headline change is clip length. Where most of the field spent the past year stuck in roughly a 4-to-15-second band, Seedance 2.5 generates a full 30-second clip in a single pass — no post-stitching, and able to carry scene changes and tempo shifts inside one continuous render. That is a meaningful jump, because a 30-second clip is long enough to be a finished short on its own rather than a fragment you have to extend.
The other upgrade is input breadth. Seedance 2.5 accepts up to 50 reference inputs in one request — images, audio, video, style references, and 3D models — up from 12 in Seedance 2.0, which gives much tighter control over look, motion, and consistency across a shot. ByteDance also describes native 4K output and post-generation editing that preserves the clip's visual style. Treat specific resolution and capability figures as a snapshot, since the model was still in beta when these were announced.
Two caveats matter. First, Seedance 2.5 is not generally available yet: ByteDance framed it as an enterprise beta with a public launch targeted for early July 2026, so access and limits will shift. Second, there is no published independent benchmark for 2.5 — any score circulating now is unverified. The predecessor, Seedance 2.0, has led the independent Artificial Analysis Video Arena for both text-to-video and image-to-video, which is the credible signal for the line's quality.
The interesting thing about Seedance 2.5 is the 30-second clip — that length is finally enough to be a short on its own, not a 5-second fragment you have to pad. But "a 30-second MP4" is still not a finished post: it has no captions for sound-off feeds, no per-platform aspect ratio, no hook text on the silent first second, and no copy around it. That gap is exactly what Kompozy closes. Drop a Seedance export in and Kompozy burns in branded, on-style captions, reframes the clip to 9:16, 1:1, or 16:9 per destination, and layers a hook overlay through HyperFrames so the autoplay opening lands — then it schedules and publishes the same short across TikTok, Reels, YouTube Shorts, Facebook, LinkedIn, X, Pinterest, and Threads from one queue.
Because the clip carries multiple beats, it also seeds a whole content unit. Feed the concept into Kompozy and it fans the same idea into a carousel breaking down each beat, a quote card pulled from the script, a blog draft, and platform-native captions written in your voice through the Persona Brief. One 30-second generation becomes a week of cross-platform posts instead of a single upload. Seedance owns the render; Kompozy owns the captions, the formats, the schedule, and the publish.
It is ByteDance's text- and image-to-video model, unveiled at the Volcano Engine FORCE conference on June 23, 2026. Its standout feature is generating a 30-second clip in a single pass without stitching, accepting up to 50 reference inputs. It powers ByteDance apps like Dreamina, Doubao, and CapCut.
Yes — that is the headline change. It generates a continuous 30-second clip in a single pass, including scene changes and tempo shifts, without post-stitching, where most models had been limited to roughly 4 to 15 seconds.
Not generally. ByteDance positioned it as an enterprise beta at launch, with a public release targeted for early July 2026. Access, limits, and pricing were still settling at the time of writing, so confirm current availability through Volcano Engine.
Seedance 2.5 adds native 30-second single-pass generation and raises reference inputs to 50 (from 12 in 2.0), with ByteDance describing native 4K output and style-preserving edits. Seedance 2.0 is the version that has topped the Artificial Analysis Video Arena; there is no published independent benchmark for 2.5 yet.
Seedance generates the clip but does not publish it. Bring the export into Kompozy to add branded captions, reframe per platform, stack a hook overlay, fan it into a carousel and captions in your voice, and schedule and publish across TikTok, Reels, YouTube Shorts, X, LinkedIn, and more from one queue.