Wan 2.1 Image-to-Video 480p

Wan 2.1 Image-to-Video 480p is an image-to-video generation model from the Wan 2.1 family, catalogued here under the WaveSpeed provider alias wavespeedai/wan-2.1-i2v-480p. Rather than generating a clip from text alone, it takes an input image as the starting frame and produces motion from it, typically guided by an accompanying text prompt describing the desired action or camera movement. The "480p" designation identifies the output resolution tier of this variant, distinguishing it from higher-resolution siblings in the same Wan 2.1 image-to-video line. Because this is a video model rather than a chat model, the specs that matter are different: there is no context window or token throughput to compare. Instead the relevant characteristics are input conditioning (a reference image plus prompt), output resolution (480p), clip duration, and generation latency and cost per clip or per second of video. The 480p tier generally trades pixel resolution for lower compute per generation, which is why it is often the variant chosen for iteration, previews, and high-volume batch work before committing to a higher-resolution render. In practice, image-to-video models like this one are used to add motion to product photos, illustrations, character art, and stills from other generative image models. Compared with text-to-video, conditioning on an image gives much tighter control over subject appearance, composition, and style, since the first frame is supplied rather than invented. We do not track every parameter this specific endpoint exposes — check the provider's API documentation for supported clip lengths, frame rates, seeds, and prompt options, and see the pricing table on this page for how each provider bills generations.