Download
Step-Video-T2V is a state-of-the-art text-to-video foundation model developed to generate videos from natural-language prompts; its 30B-parameter architecture is designed to produce coherent, temporally extended video sequences — up to around 204 frames — based on input text. Under the hood it uses a compressed latent representation (a Video-VAE) to reduce spatial and temporal redundancy, and a denoising diffusion (or similar) process over that latent space to generate smooth, plausible motion and visuals. The model handles bilingual input (e.g. English and Chinese) thanks to dual encoders, and supports end-to-end text-to-video generation without requiring external assets. Its training and generation pipeline includes techniques like flow-matching, full 3D attention for temporal consistency, and fine-tuning approaches (e.g. video-based DPO) to improve fidelity and reduce artifacts. As a result, Step-Video-T2V aims to push the frontier of open-source video generation.
Features
- Text-to-video generation: synthesizes video sequences (dozens to hundreds of frames) from natural-language prompts
- Bilingual support: accepts prompts in English or Chinese through dual text-encoders
- Compressed latent space representation (Video-VAE) for efficient spatial + temporal encoding and reduced computational load
- Full 3D attention / diffusion-based video synthesis ensuring temporal coherence and smooth motion across frames
- Built-in training and generation pipeline including flow-matching, latent-space denoising, and optimization strategies for video quality (e.g. DPO)
- Open-source release — enabling creators to experiment, fine-tune, or build on top of an end-to-end video foundation model
Project Samples
