Bailian: wan2.5-t2v-preview:free

Wan2.5-t2v-preview is a cutting-edge multimodal video generation model from Alibaba's Tongyi Wanxiang series. As a research preview, it introduces groundbreaking native audio-visual synchronization, allowing the model to generate high-fidelity video (up to 1080p) and its corresponding synchronized audio (speech, background music, or ambient sound) simultaneously from a single text prompt. It features significantly enhanced motion dynamics and prompt adherence compared to the 2.2 series, supporting various aspect ratios and complex narrative descriptions.