Overview

HunyuanVideo was released by Tencent in December 2024 as an open-weight text-to-video model with over 13 billion parameters. Its technical report described a systematic framework covering data curation, architecture and training infrastructure, and the release helped narrow the gap between open and closed video generation.

Highlights

  • Over 13 billion parameters
  • Unified architecture for image and video generation
  • Released with a detailed technical report
  • Community tools for running it with less memory
  • Image-to-video variants released later

Typical uses

Research on video generation, creative experiments and building self-hosted video pipelines.

Things to consider

  • The community license has territorial and usage restrictions; read it carefully.
  • High GPU memory requirements at full quality.
  • Generation is slow compared with image models.

Specifications are taken from the provider’s official announcement, documentation or model card. Benchmark results depend on evaluation settings; see how AI benchmarks work.