ByteDance launches Seedance 2.5 for one-take 30s audio-video generation
emmanuelvivier · x · 2026-08-04
ByteDance's Seed team has officially launched Seedance 2.5, its new-generation video creation model. Built on a unified multimodal audio-video joint-generation architecture, the major upgrades include:
- Long-form & Multi-round Extension: Generates high-quality 30-second audio-video clips in one pass with support for multi-round extensions, optimizing transitions for consistent, multi-minute videos.
- Advanced Multimodal Referencing: Accepts up to 30 images, 10 video clips, and 10 audio clips simultaneously as reference materials to capture creative intent accurately.
- Precise Editing: Offers timestamp-level control for highly stable and targeted audio-video editing.
Related event: ByteDance's Dreamina Debuts Seedance 2.5 with 30s Video and 50 References(15 posts)→
More from Multimodal
- Wan 3.0 Video Generation Demo: Single Prompt, First Attempt — OdinLovis · 2026-08-24
- Wan 3.0 launches on fal with native 30-second video generation — aziz4ai · 2026-08-24
- Animation workflow combining LTX2.3 FLFA2V and Suno Audio — JustRuss79 · 2026-08-24
- Minimax H3 T2VA supports 15+ characters simultaneously on screen — beatlepol · 2026-08-24
- How to replace an image element with Flux Kontext or Qwen Image Edit? — Wemos_D1 · 2026-08-24
- Creating Halloween-style shorts with H3's R2V template — goulash47 · 2026-08-24