Seedance 2.5

Seedance 2.5 is ByteDance Seed's latest officially released video-generation model. It expands one-pass generation to 30 seconds, accepts larger multimodal reference sets, generates synchronized audio and video, and adds multi-round extension plus more precise video-editing controls.

ByteDance announced and released Seedance 2.5 on July 31, 2026. Its launch post introduced the model and initially described BytePlus API access as forthcoming; current BytePlus documentation now lists the model for enhanced video generation.

Yes. As verified against ByteDance Seed on September 8, 2026, Seedance 2.5 is the latest officially released generation in the Seedance model family, following Seedance 2.0, which launched on February 12, 2026.

Yes. Current BytePlus enhanced video-generation documentation lists Seedance 2.5 as a supported and callable model. The documentation covers text-to-video, image-to-video, multimodal reference workflows, combined references and video editing.

The current BytePlus API model ID is dreamina-seedance-2-5-260628. Use that exact identifier when following the current BytePlus enhanced video-generation documentation, and recheck the official model list before production integration.

Current BytePlus API documentation lists a generation range of 4–30 seconds. ByteDance also describes up to 30 seconds of audio-video generation in a single pass, with multi-round extension available for longer connected sequences.

Current BytePlus enhanced-generation API documentation does not list 1080p or 4K for dreamina-seedance-2-5-260628. It lists 480p and 720p. That API-specific fact does not establish a permanent limit for every consumer surface or future deployment.

The current BytePlus enhanced video-generation documentation lists 480p and 720p output for dreamina-seedance-2-5-260628, at 24 fps. Resolution availability is an API specification and should not be treated as a direct measure of generation quality.

Yes. ByteDance describes Seedance 2.5 as generating synchronized audio and video in one pass. The model can also accept audio references as part of a multimodal reference workflow, subject to the documented input limits.

Seedance 2.5 can use up to 30 reference images in one generation workflow, according to ByteDance’s official launch announcement. Images form one part of the model’s maximum 50-asset multimodal reference budget.

Yes. ByteDance says Seedance 2.5 accepts up to 10 reference video clips in a single workflow. Those clips can be combined with image and audio references within the model’s published multimodal limits.

Yes. Seedance 2.5 accepts up to 10 reference audio clips according to ByteDance’s launch material. Audio references may be combined with up to 30 images and 10 videos in one multimodal workflow.

Yes. Official materials describe reference-based editing, upgraded green-screen and camera-perspective controls, and timestamp-level editing. BytePlus also documents video editing as one of the supported enhanced-generation workflows for the model.

Seedance 2.5 doubles one-pass duration from 15 to 30 seconds, raises the reference budget from 9/3/3 to 30/10/10 images/videos/audio, and adds stronger extension and timestamp-level editing. Current API resolution listings differ by model.

BytePlus documents API access for developers, while ByteDance’s own model pages describe official capabilities. Seedance2Video currently exposes only the models shown in its live selector; if 2.5 is absent there, this page is a reference hub.

No. Vividra Labs LLC operates seedance2-video.com independently, and the site is not affiliated with ByteDance. ByteDance Seed and BytePlus are the primary sources for model identity, capabilities and API specifications.

There is no single universal free-access answer. ByteDance product access and pricing are governed by the applicable ByteDance or BytePlus terms. Seedance2Video has separate credit pricing for the models currently shown in its generator; see its live pricing page.

Commercial use depends on the terms of the specific provider and access surface. Review current ByteDance or BytePlus terms for their product. Seedance2Video’s separate Terms govern outputs made through models actually available on this independent site.