Wan, developed by Alibaba and released under an open Apache 2.0 license in 2025, joins Genmo, also covered in this directory, as one of the relatively few genuinely open-source models in this category, with model weights and inference code available for teams that want to run or customize it directly rather than depending on a closed vendor service.
Its more distinctive technical capability is generating synchronized audio and video together in a single step, voice, sound effects, and lip-sync produced alongside the visual content rather than layered on afterward in a separate process, backed by Alibaba's substantial cloud and AI research infrastructure.
Wan is free to use, backed by Alibaba's cloud infrastructure. For a technical team that wants an open-source model generating synchronized audio and video together in one step, backed by major-platform resources rather than an independent research lab, Wan's open license and audio-video synchronization address that combination directly.
