
10 billion parameters. Premium 3D from text or images. PBR materials, pose control, multi-format export.
Rodin is the largest AI 3D generation model available, with 10 billion parameters trained on an extensive dataset of 3D assets. It produces the highest fidelity geometry and PBR materials from text prompts, single images, or up to 5 reference images. Supports T-pose and A-pose output for character pipelines. Exports to GLB, FBX, OBJ, USDZ, and STL.

“A medieval sallet helmet displayed on dark velvet in a museum hall, aged steel with gold filigree trim along the visor edge, tapestries in the background, artefact photography”

“An angular supercar clay model on a turntable in a design studio, matte grey finish, sharp creases and aggressive front splitter, sketch boards on the wall behind, automotive concept photography”

“An ornate Victorian street lamp with griffin sculptures and scrollwork on the column, weathered green patina, frosted lantern glowing at dusk on a cobblestone street, heritage architecture”

“A white anatomical human skull on dark slate, three-quarter view showing cranial sutures, teeth, and orbital detail, museum label in the corner, medical reference photography”

“A Bengal tiger taxidermy specimen in a natural history museum diorama, painted forest backdrop, realistic fur and stripe detail, warm display lighting”

“A sci-fi armour chest piece with brushed steel plating and glowing blue energy conduits, displayed on a plinth in a futuristic lab, holographic schematics in the background, concept art render”
Type a text description of the 3D model you need, or upload reference images from 1-4 angles.
Pick a 3D generation model and configure settings. See the credit cost before you generate.
Your 3D model is ready in seconds. Download in FBX, GLB, OBJ, or other formats for your pipeline.
Jump into the Studio and start generating. Plans from £10/month.
Choose a PlanRodin is built by Hyper3D with 10 billion parameters, making it the largest publicly available AI 3D generation model. The scale translates directly to output quality: sharper edges, more accurate proportions, better material separation, and fewer artefacts than smaller models. For hero assets, product visualisation, and any use case where quality matters more than speed, Rodin is the tool.
Input flexibility covers every workflow. Generate from a text description for early concept exploration. Upload a single reference image for faithful reconstruction. Provide up to 5 images from different angles and Rodin fuses them into a geometrically accurate model. The multi-image fusion mode cross-references perspectives for precise proportions.
Character workflows benefit from T-pose and A-pose output options. Generate a character in a standard rigging pose, ready for your skeleton and animation pipeline. Bounding box controls let you specify the exact dimensions of the output in world units. PBR materials (base colour, metallic, roughness, normal) are generated by default.
The largest AI 3D model available. More parameters means better understanding of 3D geometry, materials, and spatial relationships. The quality gap over smaller models is most visible on complex surfaces, thin structures, and material transitions.
Upload up to 5 reference images. Rodin cross-references angles and fuses them into a single accurate model. No photogrammetry setup required. Works with photos, renders, or concept art.
Generate characters in T-pose or A-pose for direct import into rigging pipelines. Specify bounding box dimensions for accurate world-scale output.
GLB for web and AR. FBX for Maya and game engines. OBJ for legacy tools. USDZ for Apple AR. STL for 3D printing. Every format includes PBR materials where supported.
A small indie studio building creative tools the way they should be built. No VC theatre, no funnel games, no faceless support.