Features
- Text-to-video generation with realistic motion and strong prompt adherence
- Open-source 10B-parameter AsymmDiT model under Apache 2.0
- AsymmVAE video compression model
- Runs locally with multi-GPU support
- LoRA fine-tuning to customize the model
- Gradio interface and command-line generation
- API integration and ComfyUI support
Pricing Details
Mochi 1 (Genmo AI) is available for free and is open source; it does not rely on a subscription model or paid plans. The project provides users with access to the source code, model weights, and runtime tools under the Apache 2.0 license, allowing it to be used, developed, and modified within various projects. The model can be run locally after meeting the necessary technical requirements, or integrated into other environments and tools such as ComfyUI or via APIs. There are currently no official paid plans associated with Mochi 1, as the model is provided entirely as an open-source project for developers and researchers.