Seedance 2.0 is ByteDance's flagship video generation model, the successor to Seedance 1.0 Pro. It utilizes a unified multimodal audio-video architecture that accepts text, image, audio, and video inputs, supporting text-to-video, image-to-video, and reference-to-video generation. The model produces 60-90 second clips at up to 1080p resolution, featuring multi-shot cuts, realistic physics, and natively synchronized audio in a single pass. A key strength is character consistency, which preserves identity across occlusions and scene transitions, using multi-image references to maintain style and composition. Seedance 2.0 is designed for cinematic production, episodic content, and brand advertising.