Shengshu Tech launched Vidu S2, featuring S2-Avatar for interactive digital humans and S2-Editing for real-time video modification. Key innovations include Self-Replay Forcing training and VR spatial video support. S2-Avatar offers 720P resolution with dynamic element integration, while S2-Editing enables style transfer, virtual try-on, character and background replacement. The models lead in StreamAV-Bench and Sparkle-Bench evaluations, demonstrating superior audio-video sync and consistency. Future focus remains on reducing latency and enhancing panoramic spatial video quality for VR headsets.