Choose Auto, TokenLab Verified, or Official for each request, with prices shown up front. See what's new

Hyper3D: Rodin Gen-2

Rodin Gen-2 is a 10 billion parameter 3D generation model from Hyper3D built on the BANG architecture. It creates production-ready assets from text prompts or up to 5 input images, with 4x geometry improvement over Gen-1. It supports quad-dominant and raw mesh modes, PBR and shaded materials, configurable polygon counts, T/A pose control, bounding box constraints, and output in GLB, FBX, OBJ, USDZ, and STL formats.
Compare models
rodin-gen-2
AvailableHyper3D3DAsync
Price
From $0.40
Modalities
3D

About Rodin Gen-2

Rodin Gen-2 is a 10 billion parameter 3D generation model from Hyper3D, built on the BANG architecture. It creates assets from a text prompt or up to five images and outputs meshes with PBR or shaded materials, in GLB, FBX, OBJ, USDZ, or STL. It offers more control over topology and posing than Hyper3D Gen2.

Where it works well

  • Recursive part-based generation splits objects into coherent components, which suits complex, structured assets.
  • Hyper3D reports a fourfold improvement in geometry quality over Gen-1.
  • Mesh type is selectable between quad-dominant and raw triangle output, with a configurable polygon count.
  • Controls such as T or A pose for characters and bounding-box constraints steer the result's proportions.
  • Baked normals can carry high-polygon detail onto a low-polygon model for real-time use.

When to choose another model

  • It produces static meshes; rigging, animation, and retopology fine-tuning continue in a DCC tool.
  • Multi-image input stops at five images, so large photo sets for photogrammetry-style capture are out of scope.

Getting started

  1. Create API key

    Create a key in Console, then use it with every model on the platform.

  2. Send your first request

    Copy the example for your language and run it against the endpoint.

    Image to 3DText to 3DTokenLab endpoint
    POST/v1/3d/generations
    curl -X POST "https://api.tokenlab.sh/v1/3d/generations" \
      -H "Authorization: Bearer sk-xxx" \
      -H "Content-Type: application/json" \
      -d '{
      "model": "rodin-gen-2",
      "prompt": "A stylized ceramic astronaut figurine."
    }'

Parameters

Image to 3D

ParameterValues and limitsDefaultNotes
formatglbglbNot provided
imageInputs: ≤ 1Not providedLimits apply when this input is supplied. The public API accepts one image string. Multiple image arrays are not exposed.
image_urlInputs: ≤ 1Not providedLimits apply when this input is supplied. The public API accepts one image string. Multiple image arrays are not exposed.
n11Use an integer.
operationRequiredNot providedNot providedNot provided
output_formatglbglbNot provided
promptNot providedNot providedMinimum length: 2 characters. Maximum length: 1024 characters.
seed≥ 0, ≤ 65535Not providedUse an integer.

Text to 3D

ParameterValues and limitsDefaultNotes
formatglbglbNot provided
imageInputs: ≤ 1Not providedLimits apply when this input is supplied. The public API accepts one image string. Multiple image arrays are not exposed.
image_urlInputs: ≤ 1Not providedLimits apply when this input is supplied. The public API accepts one image string. Multiple image arrays are not exposed.
n11Use an integer.
operationRequiredNot providedNot providedNot provided
output_formatglbglbNot provided
promptNot providedNot providedMinimum length: 2 characters. Maximum length: 1024 characters.
seed≥ 0, ≤ 65535Not providedUse an integer.

Fields outside this contract are rejected.

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Image to 3D

per request
Official price
$0.40
Verified price
—
Discount
—

Text to 3D

per request
Official price
$0.40
Verified price
—
Discount
—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours
Requests
Success rate
P95 latency
Total tokens
Model performance
Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Rodin Gen-2 in Console with a prompt ready to edit or send.

Help me create with rodin-gen-2 using text-to-3d at /v1/3d/generations. Show the result, status, and cost.

Use cases

  • Game-ready props

    Request a quad-dominant mesh at a target polygon count, with baked normals, and drop it into an engine as a starting asset.

  • Character base meshes

    Generate a humanoid in a T or A pose so it is ready for rigging, then refine the model in a sculpting tool.

  • 3D printing models

    Export STL from a text prompt or photos of an object, then adjust wall thickness and supports in a slicer.

Prompt examples

A medieval knight in a T pose, plate armor with a tabard, neutral expression, low-poly game-ready quad mesh.

A mechanical pocket watch with an open case, visible gear train, brass and steel PBR materials.

Combine the five attached views of the sneaker into one 3D model, keeping the stitching and sole pattern.

FAQ

What formats does Rodin Gen-2 export?

GLB, FBX, OBJ, USDZ, and STL. That covers game engines, DCC tools, AR viewers on Apple devices, and 3D printers, so you rarely need an extra conversion step.

How many images can I use?

Up to five input images, or a text prompt alone. Several views of the same object help the model reconstruct sides that a single photo does not show.

What is the BANG architecture?

It is the design Hyper3D uses for Gen-2, trained at 10 billion parameters. It generates objects recursively by parts, so complex items are divided into components rather than made as a single blob.

Can I control the polygon count?

Yes. Rodin Gen-2 supports configurable polygon counts and a choice between quad-dominant and raw mesh modes, along with PBR or shaded materials, so you can match the asset to a real-time or offline target.

How much does Rodin Gen-2 cost?

On TokenLab, Rodin Gen-2 costs $0.40 per request. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

Which endpoint should Rodin Gen-2 use?

Use https://api.tokenlab.sh/v1/3d/generations for Rodin Gen-2. The request example below shows the matching code shape.

Which operations does Rodin Gen-2 support?

Rodin Gen-2 supports Image to 3D, Text to 3D. Select an operation above to see its endpoint and request example.

Compare Rodin Gen-2

Sources

Reviewed Oct 2, 2026

More from Hyper3D

Related models