Seedance 2.0 Mini

• 🏢 ByteDance Seed's video generation family, text-to-video variant.
• 🆕 Newest Seedance tier, released June 2026.
• 🎬 Cinematic output with native audio-video joint generation.
• 🔧 Unified multimodal model accepting text, image, audio, video.
• 📏 Multi-shot cinematic clips across multiple aspect ratios.
• 🎯 Director-level camera, lighting, and motion control.
• ⚡ Improved motion and physics for high-action sequences.
• 👁️ Strong scene and character consistency across frames.

ByteDance

ByteDance is a Chinese internet technology company headquartered in Beijing, widely known as the parent company behind TikTok and Douyin. Beyond social media, ByteDance has invested heavily in artificial intelligence research, building generative media models…

Seedance 2.0 Mini is the text-to-video member of ByteDance's Seedance 2.0 video family, developed by the company's Seed research division. Like its siblings, it is built on a unified multimodal architecture that accepts text, images, audio, and video as inputs and produces audio-synchronized cinematic video, with multi-shot generation, real-world physics, and director-level camera control described for the Seedance 2.0 platform.

The "Mini" tier sits alongside the standard Seedance 2.0, the higher-quality Seedance 2.0 Enhanced, and the latency-oriented Seedance 2.0 Fast, and it ships with image-to-video and reference-to-video companions, Seedance 2.0 Mini and Seedance 2.0 Mini R2V.

Against the prior Seedance 1.5 Pro generation, ByteDance describes the 2.0 line as adding native audio-video joint generation, improved motion and physics for complex interactions like sports, dancing, and collisions, and significantly stronger consistency for faces, clothing, text, scenes, and styles. The family is positioned around reference-driven control and continuing scenes with consistent characters and camera work.

Specific parameter counts, licensing, and benchmark figures for the Mini variant were not available in primary ByteDance documentation at the time of writing, so capability details here reflect the shared Seedance 2.0 family description rather than tier-specific measurements.