Hourly · 2026-06-23 16:00
ByteDance Seedance 2.5 Breaks the 30-Second AI Video Barrier
ByteDance unveils Seedance 2.5 at Volcano Engine FORCE conference — single clips up to 30 seconds, 4K, scene changes, no stitching.
ByteDance introduced five new AI models at Volcano Engine's FORCE conference, headlined by Seedance 2.5. The video model generates single clips up to 30 seconds without post-stitching, with scene changes and tempo shifts built in.
It can process up to 50 simultaneous inputs — reference images, audio, and more — making it viable for multi-character film scenes. Users can also edit generated videos after the fact while preserving visual style.
Seedance 2.0 simultaneously gets a 4K upgrade with 10-bit color. The broader lineup includes Doubao 2.1 Pro (claimed to cost 80% less than Claude Opus 4.6), Seedream 5.0 Pro for images, and Seed-Audio 1.0. Seedance 2.5 launches in early July.
Sources: The Decoder, The Information
