Genmo

Genmo is the AI video lab behind Mochi 1, an open-source text-to-video model released with weights on Hugging Face and a companion hosted playground at genmo.ai. The site pitches the model as a state-of-the-art open alternative for turning written prompts into short generative video, with an emphasis on physical-world understanding and motion coherence rather than still-frame animation.

What makes Genmo notable is the openness angle: unlike Runway, Pika, or Sora, the underlying weights are downloadable and can be run locally on sufficiently beefy GPUs, which is attractive to researchers, indie tool builders, and studios that need to keep prompts and outputs in-house. The playground offers a free way to try it in the browser without provisioning hardware, though heavy generation is naturally gated. Pricing on the marketing page is thin, so expect the hosted side to move toward credits or a paid tier as the product matures.

It sits in the same conversation as Open-Sora and CogVideoX for teams evaluating open video generators, and pairs well with an image model plus an upscaler if you want longer, higher-resolution finals. Clip length, resolution, and fine control still lag closed leaders, so treat it as a rapidly moving research-grade tool rather than a finished production pipeline.