ByteDance: Seedance 2.0 Fast

ByteDance: Seedance 2.0 Fast is a video generation model from ByteDance, the company behind the Seedance line of generative video systems. It sits in the video category and is positioned as the faster, lighter option in the Seedance 2.0 family, aimed at workloads where generation speed matters more than maximum fidelity. It was released by the provider on 15 April 2026.

The model accepts text, image, video and audio as inputs, and produces video as its only output. It handles text-to-video, image-to-video with first and last frame control (the frame images are named first_frame and last_frame), and multimodal reference-to-video, where you can mix reference material across the supported input types. Output video can be generated at 480p or 720p, in aspect ratios of 1:1, 3:4, 9:16, 4:3, 16:9, 21:9 and 9:21, with durations selectable from 4 to 15 seconds in one second steps. Audio generation is available alongside the video. No context length is published for this model, so token based planning does not apply here.

Those capabilities point to a fairly specific set of jobs. Text-to-video and image-to-video cover short ad spots, social clips and vertical formats such as 9:16 and 9:21, while the wider 21:9 and 16:9 ratios suit trailer or banner style output. First and last frame control is useful when you need a shot to start and end on defined images, which fits product shots, transitions and storyboard driven work. Reference-to-video with video and audio inputs helps when you want the output to follow the look, motion or sound of existing material rather than a written description alone. Since durations run from 4 to 15 seconds, the model fits short form pieces rather than long sequences.

You call it through api14's OpenAI-compatible API using the model id "bytedance/seedance-2.0-fast". The request format follows the same structure as other models on the gateway, so existing OpenAI style clients work without special handling. Among the listed parameters, the model supports frequency_penalty. Other settings such as aspect ratio, resolution, duration, first and last frame images, and audio generation are part of the model's own generation options rather than the shared parameter list.

Because the model is built for speed and lower cost, it can be the wrong pick when a shot needs the highest possible visual fidelity, since a heavier Seedance variant would be the natural comparison. No context length, maximum output token count or broader parameter set is published here, so if your pipeline depends on sampling controls beyond frequency_penalty, check what the gateway currently exposes for this model id before committing to it.