Seedance 2.5 US

• 🏢 ByteDance's flagship video family, US-region text-to-video endpoint
• 📏 Up to 30-second audio-video clips generated in one pass
• 🆕 Multi-round extension for building longer continuous sequences
• 🔧 Timestamp-level editing of audio and video segments
• 👁️ Improved shot transitions and scene changes for continuity
• 🎯 Clay-render, motion and creative referencing for spatial control
• 🧠 ByteDance reports gains in image, audio and motion quality

ByteDance

ByteDance is a Chinese internet technology company headquartered in Beijing's Haidian district, with an associated variable-interest entity incorporated in the Cayman Islands. Best known as a consumer internet operator at enormous scale, it has also become a…

Seedance 2.5 US is the United States regional endpoint of ByteDance's flagship Seedance video family, exposed here as a text-to-video route. It sits alongside Seedance 2.5 US for image conditioning and Seedance 2.5 US R2V for reference-driven generation, and mirrors the earlier general release of Seedance 2.5.

Against its own predecessors, the generational gains ByteDance describes are concrete. The company presents Seedance 2.5 as producing high-quality 30-second audio-video clips in a single pass, with multi-round extension for longer pieces, improved shot transitions and scene changes for continuity across cuts, and gains in image, audio and motion quality relative to the earlier Seedance generations. That positions it as a step up from the shipping Seedance 2.0 line and the earlier Seedance 1.5 Pro within this catalog's lineage.

Control is the other axis of improvement. According to ByteDance, the model adds timestamp-level editing of audio and video segments, plus clay-render, motion and creative referencing so creators can specify spatial structure, character poses, motion paths and camera angles rather than relying on prompt wording alone. ByteDance also reports systematic work on textures, skin and eye detail, lighting and colour saturation to reduce the artificial look common in generated footage.

Within the Seed family it shares a lineage with the Seedream image models, including Seedream V5 Pro. Published capability details for this release come from ByteDance's own Seedance 2.5 launch blog and model page.