ByteDance's Seedance 2.5 Generates 30-Second AI Video Clips with Scene Changes and 50+ Simultaneous Inputs

ByteDance unveiled Seedance 2.5 at Volcano Engine's FORCE conference, with a launch slated for early July. The model generates single video clips up to 30 seconds long — without post-stitching — including scene changes and tempo shifts. It accepts up to 50 simultaneous reference inputs such as images and audio, making it suited for complex multi-character film scenes, and allows post-generation edits that preserve the original visual style.

Alongside the flagship announcement, ByteDance upgraded Seedance 2.0 to native 4K with 10-bit color depth. The company also introduced three additional models through its Volcano Engine cloud business: the Doubao 2.1 Pro language model (priced roughly 80% below Claude Opus 4.6, per the company), the Seedream 5.0 Pro image model, and the Seed-Audio 1.0 audio model.