Summary
Around August 7, 2026 ByteDance made Seedance 2.5 fully available via its Experience Center and API. The model generates up to 30 seconds of native single-shot video with synchronized audio, accepts up to 50 multimodal references (images, video, and audio), and speaks 10+ languages, while the parallel Seedance 2.0 was upgraded to 4K output.
What changed
Seedance 2.5 became generally available through ByteDance's Experience Center and API. It produces up to 30 seconds of native single-shot video, co-generates audio and video in a shared latent space, accepts up to 50 multimodal references (30 images, 10 video, 10 audio), and supports native audio in 10+ languages. Seedance 2.0 was simultaneously upgraded from a 720p/1080p ceiling to 4K. Access requires an account balance above roughly USD 30 or an active Seedance 2.0 resource package; there is no free quota.
Why it matters
Joint audio-video generation in a single pass, at 30-second lengths with strong character and motion consistency, pushes the frontier of generative media and pressures rivals like OpenAI Sora and Google Veo. Native multilingual audio and multimodal referencing make it directly usable for ads, localization, and product content, and API availability turns it into infrastructure other apps and agents can call.
Evidence excerpt
Seedance 2.5 generates up to 30 seconds of native single-shot video, accepts up to 50 multimodal references (30 images, 10 video, 10 audio), adds native audio with 10+ languages, and co-processes visual and audio signals in the same latent space; Experience Center and API were fully available around August 7, 2026.