Alibaba's Tongyi Wanxiang team has open-sourced Wan-Animate-2, an end-to-end character animation framework that streams at 24 frames per second in real-time without skeletal pose extraction. The architecture feeds driving video directly into a diffusion Transformer, skipping intermediate pose-estimation and motion feature compression. Head-to-head tests show results competitive with closed-source commercial leaders such as ByteDance's Dreamina and Kuaishou's KLING MotionControl. The framework produces cleaner hand deformation and fewer limb artifacts on challenging motions. It supports multi-character and multi-camera scenarios, allowing a single driving video to drive multiple target characters simultaneously and rendering from multiple explicit camera poses without retraining. The system also handles non-standard aspect ratios and varied character body types. Wan-Animate-2 closes the gap in character animation, which had been a bottleneck in commercial-grade AI video, and its weights can be run locally by developers. The release follows Wan 3.0, released earlier in August, which opened document-input video generation for enterprise use. Together, the two releases cover both ends of the production stack. The article notes that Chinese labs already match closed-source commercial quality, and the remaining question is how fast downstream tooling integrates the new model.