ByteDance has launched its new video model Dreamina Seedance 2.5 and its genuinely impressive.
Rohan Paul Twitter · Rohan Paul (@rohanpaul_ai) · 2026-08-03
ByteDance launches Dreamina Seedance 2.5, a video generation model that accepts up to 50 reference files and produces 30 seconds of continuous video, with timestamp-based prompting enabling shot-by-shot directorial control.
Appears in
Extraction
Topics: video-generationmultimodal-aigenerative-videobytedance
Claims
- Seedance 2.5 accepts up to 50 reference files simultaneously — up to 30 images, 10 videos, and 10 audio clips — for a single generated clip.
- The model produces up to 30 seconds of continuous video in a single generation pass.
- Timestamp-based prompting lets creators specify different actions at precise time intervals, functioning as an in-prompt shot list.
- Seedance 2.5 launches exclusively on the Dreamina platform for its first seven days as a world-first global release.
Key quotes
AI video has been mostly a guessing game until now: somebody types a long paragraph and hopes the model imagines the right face, the right room, the right camera move.
a creator can write [0s-5s] for one action and [5s-10s] for the next and the model will use that as a shot list.
Use up to 50 reference files for a single clip and produce 30 seconds of continuous video in one take.