microsoft/trellis2
TRELLIS.2 is a state-of-the-art 4B parameter 3D generative model for high-fidelity image-to-3D generation with complex topologies, sharp features, and full PBR materials.
Model Inputs
Input image to generate 3D asset from
Number of sampling steps (more steps = better quality, slower)
Pipeline resolution for quality/speed tradeoff
Texture size for GLB export
Random seed. Leave blank to randomize the seed
Result
Pricing
microsoft/
trellis2
Pricing for Synexa AI models works differently from other providers. Instead of being billed by time, you are billed by input and output, making pricing more predictable.
For example, generating 100 requests costs about $25.00.
Check out our docs for more information about how per-request pricing works on Synexa.
| Provider | Price ($) | Saving (%) |
|---|---|---|
| Synexa | $0.25 | - |
| replicate | $0.82 | 69.5% |
Readme
TRELLIS.2 is a state-of-the-art large 3D generative model (4B parameters) designed for high-fidelity image-to-3D generation. It leverages a novel "field-free" sparse voxel structure termed O-Voxel to reconstruct and generate arbitrary 3D assets with complex topologies, sharp features, and full PBR materials.
Key features:
- 4B parameter model for high-quality 3D asset generation from single images
- Uses O-Voxel (field-free sparse voxel structure) for representing complex topologies
- Supports multiple pipeline resolutions (512, 1024_cascade) for quality/speed tradeoffs
- Generates textured GLB models with configurable texture sizes
- Automatic background removal and image preprocessing
- Full PBR (Physically Based Rendering) material support
Usage notes:
- Higher resolution (1024_cascade) produces better quality but takes longer
- More sampling steps improve quality at the cost of speed
- Texture size affects the detail level of the exported GLB model
- The model automatically preprocesses input images (removes background and crops)