Skip to main content
Back to timeline
Vercel NewsSource publication:

Seedance 2.5 Flash lands on AI Gateway: faster and cheaper, with output close to the standard model

Synopsis

ByteDance's Seedance 2.5 Flash is now available on AI Gateway as a faster, lower-cost version of Seedance 2.5. It generates video with synchronized audio from text prompts or image, video, and audio references, with clips of 4 to 30 seconds at up to 720p and 24 frames per second in landscape, square, or portrait, and it can also edit or extend existing videos. ByteDance says its output is close to the standard model in human realism, reference-based editing, and prompt adherence.

Source-provided article image: Seedance 2.5 Flash now available on AI Gateway
Vercel

Interpretation

ByteDance's Seedance 2.5 Flash is live on AI Gateway, and developers call it through the model identifier bytedance/seedance-2.5-flash to generate video with synchronized audio. Flash is positioned as a faster, lower-cost version of Seedance 2.5, and the stated comparison between the two is about output quality. The stated comparison is output quality: Flash's output is close to the standard model in human realism, reference-based editing, and prompt adherence.

Generation can start from a text prompt or take images, video, and audio as reference inputs, with reference material numbered separately in the prompt as [Image 1], [Video 1], and so on. References are passed through inputReferences, and images and videos are numbered separately, so one prompt can say both "the lion from [Image 1]" and "filmed like [Video 1]." The official example calls experimental_generateVideo in AI SDK 7 with a 16:9 aspect ratio, a 5-second duration, and one JPEG image plus one MP4 video as references.

Clips can be set from 4 to 30 seconds at up to 720p and 24 frames per second, in landscape, square, or portrait, and the model can also edit or extend existing videos or take images to set the first and last frames. These are the generation parameter ranges published at Flash's launch, covering short clips up to 30 seconds and three aspect formats. The official example sets aspectRatio to 16:9 and duration to 5, showing that the parameters are specified directly in the call.

Perspective

The report is limited to what the AI Gateway launch announcement states: the model identifier, generation parameter ranges, reference input method, and the official qualitative comparison of output quality; the announcement gives no benchmark scores or third-party evaluation.

"Output close to the standard model" is the official qualitative statement across human realism, reference-based editing, and prompt adherence; the announcement provides no quantitative comparison and does not say under which conditions Flash differs from the standard version.