Published: November 5, 2026
3D Model
Hyper3D Rodin v2.5 is the full-quality version of Rodin, next to the Rodin v2.5 Fast model already in the Toolkit.
Fast caps meshes at 20K triangles; this one goes up to 1M triangles with 4K textures on High, which puts it at the top of Artlist’s 3D range alongside Hunyuan 3D v3.1 Pro.
Users pick Low / Medium / High; detail tier, mesh size, material and texture quality are fixed per preset.
Key Features
Generation
- Image-to-3D: One image becomes a full model; unseen sides are inferred
- Text-to-3D: Describe one object and get a textured model back
Output
- Dense Meshes: 50K, 500K and 1M triangles across the three presets
- HD Textures: High adds HD texture and 4K texture maps for close-up shots
- PBR Materials: On Medium and High, materials respond to scene lighting
Technical Capabilities
Details |
|
|---|---|
| Quality Presets | Low (50K triangles, shaded), Medium (500K, PBR), High (1M, PBR, HD 4K textures) |
| User Controls | Preset only; all other settings fixed by Artlist |
| Input Modalities | Image-to-3D, Text-to-3D |
| Image Input | One image |
| Text Input | Prompt up to 1,024 characters |
| Mesh / Output | Textured, triangles, GLB |
Credit Cost
Credits are charged per generation, and text-to-3D and image-to-3D cost the same.
| Mode · Preset | Credits per generation |
|---|---|
| Image-to-3D · Low, Medium | 740 |
| Image-to-3D · High | 2,210 |
| Text-to-3D · Low, Medium | 740 |
| Text-to-3D · High | 2,210 |
Hyper3D Rodin v2.5 isn’t included in any Unlimited plan, so every generation uses credits. Credit costs can change; check the model in the Toolkit for the current price.
Limitations
- Preset only, and a random seed every run: a result can’t be reproduced or fine-tuned
- On Image-to-3D the prompt is not used; the image alone drives the result
- One image in, so the back and underside are inferred
- High files are heavy at 1M triangles; use Low or Medium for real-time or web use
- Low uses a shaded material with no PBR, so it reads flat under scene lighting
Prompting Tips
- One object, clear view, plain background, filling most of the frame
- Use High for hero assets seen up close; Medium is enough for most scenes
- For characters, a neutral pose with arms away from the body gives a cleaner mesh
- In text prompts, name the object, material and style; keep it to one object