Wan 2.5Generate synchronized AI videos from text prompts and images with audio-aware output.

Wan 2.5 is a revolutionary native multimodal video generation platform that supports unified text, image, video, and audio generation. The platform features synchronized A/V output, cinematic 1080p HD quality, and human preference alignment through advanced RLHF training. Wan 2.5 uses a native multimodal architecture with a mixture of expert models to generate high-fidelity video with synchronized audio, including vocals, sound effects, and music. The platform allows users to create synchronized AI videos from text prompts and images, with audio-aware output. It is designed to be used by creators worldwide and offers several features, including precision image editing, text-to-image generation, and the ability to export 1080p HD cinematic videos. Wan 2.5 also boasts improved generation speed, video quality, and semantic compliance compared to its predecessor Wan2.2.