Member-only story
Microsoft TRELLIS: Changing How 3D Assets Are Made
A Faster, More Creative Approach to Digital Content Creation
Microsoft TRELLIS is offering a new approach to 3D asset creation, arriving just ahead of the highly anticipated OpenAI 12 update. OpenAI has been working on similar capabilities since last year and showcased examples earlier this year. It’s great to see Microsoft contributing to open-source innovation in this space.
Powered by advanced AI, TRELLIS enables users to create high-quality, customizable 3D objects effortlessly using simple text or image prompts. This innovation promises to improve 3D design workflows, making it accessible to professionals and beginner alike. Here are some examples:

What is Microsoft TRELLIS?

Microsoft TRELLIS is an AI-driven model that transforms basic text instructions into detailed 3D assets. It supports various formats — Radiance Fields, 3D Gaussians, and meshes — making it suitable for game developers, architectural visualizers, and others looking for flexibility in their workflows.
Key Features and Capabilities
- High-Quality Outputs
The outputs delivered by Microsoft TRELLIS maintain detailed shape and texture accuracy. This benefits professionals who need assets that look and feel realistic. - Flexible Input Options
Microsoft TRELLIS interprets both text and image prompts. Whether starting with a written description or using a reference image, this model can produce a fitting 3D representation. This method saves time and simplifies the creative process, benefiting both seasoned specialists and beginners. - Editing and Variations
Beyond initial generation, Microsoft TRELLIS supports easy adjustments. Users can fine-tune results, create multiple variants, or adjust specific parts of the 3D asset. This interactive approach helps teams refine outputs quickly. - Pre-Trained Models
To assist users, Microsoft TRELLIS provides large-scale pre-trained models, containing up to 2 billion parameters. These have…