Tencent Hunyuan 3D
Tencent Hunyuan 3D API reference. Generate production-ready 3D models with Pro and Rapid editions. Text-to-3D, image-to-3D, multi-view input with GLB output and optional PBR textures.
Pro Edition
POST/v1/3d-models/tencent/generate/pro/
High-quality 3D generation with advanced controls. Supports text-to-3D, image-to-3D, multi-view image input, and multiple generation types. Base cost is 60 credits (+20 for PBR, +20 for multi-view).
curl -X POST https://api.3daistudio.com/v1/3d-models/tencent/generate/pro/ \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "3.0",
"prompt": "a medieval sword with ornate handle",
"enable_pbr": true,
"face_count": 500000,
"generate_type": "Normal"
}'{
"task_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
"created_at": "2026-02-25T12:00:00Z"
}Pro Edition Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
| model | string | Yes | Model version: "3.0" or "3.1". |
| prompt | string | Conditional | Text description (max 1024 chars). Required if no image provided. |
| image | string | Conditional | Base64-encoded image (data:image/...;base64,...). Required if no prompt. |
| enable_pbr | boolean | No | Enable PBR textures. Default: false. Adds 20 credits. |
| face_count | integer | No | Target polygon count (40,000 to 1,500,000). Default: 500,000. |
| generate_type | string | No | Generation mode (see table below). Default: "Normal". LowPoly and Sketch are not available on model 3.1. |
| polygon_type | string | No | "triangle" or "quadrilateral". Only used with LowPoly. Default: "triangle". |
| multi_view_images | array | No | Array of multi-view images. Must include a "front" view. Adds 20 credits. |
Generation Types
The generate_type parameter controls what kind of output the model produces.
| Type | Description |
|---|---|
| Normal | Standard generation. Produces a textured 3D model with full geometry and materials. |
| LowPoly | Generates a model with intelligent polygon reduction applied. Use polygon_type to choose between triangle or quadrilateral topology. Not available on model 3.1. |
| Geometry | Generates a geometry only model without any textures (white model). The enable_pbr parameter has no effect in this mode. |
| Sketch | Generates a model from a sketch or line drawing input. This is the only mode where both a prompt and an image can be provided together. Not available on model 3.1. |
Multi-View Images
For Pro edition, you can provide multiple reference images from different angles. Each entry has a view_type and a base64-encoded view_image. A "front" view is always required. Model 3.0 supports: front, left, right, back. Model 3.1 adds: top, bottom, left_front, right_front.
{
"model": "3.0",
"multi_view_images": [
{ "view_type": "front", "view_image": "data:image/png;base64,..." },
{ "view_type": "left", "view_image": "data:image/png;base64,..." },
{ "view_type": "right", "view_image": "data:image/png;base64,..." }
],
"enable_pbr": true
}Rapid Edition
POST/v1/3d-models/tencent/generate/rapid/
Faster 3D generation with simplified parameters. Supports text-to-3D and image-to-3D. Base cost is 35 credits (+20 for PBR).
curl -X POST https://api.3daistudio.com/v1/3d-models/tencent/generate/rapid/ \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "a red sports car",
"enable_pbr": false,
"enable_geometry": true
}'{
"task_id": "f7e8d9c0-b1a2-3456-7890-abcdef012345",
"created_at": "2026-02-25T12:00:00Z"
}Rapid Edition Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
| prompt | string | Conditional | Text description (max 1024 chars). Required if no image provided. |
| image | string | Conditional | Base64-encoded image. Required if no prompt. |
| enable_pbr | boolean | No | Enable PBR textures. Default: false. Adds 20 credits. |
| enable_geometry | boolean | No | Enable geometry optimization. Default: false. |
Output Formats
The Pro and Rapid editions return different file formats. This is important to understand when integrating the API into your pipeline.
| Edition | Output Format | asset_type | Description |
|---|---|---|---|
| Pro | GLB | 3D_MODEL | A single .glb file with embedded textures and materials. Ready to use in most 3D applications. |
| Rapid | ZIP (OBJ) | ARCHIVE | A .zip archive containing an OBJ model with its MTL and texture files. Extract the archive to access the model. |
Need a Different Format?
If you need your model in a specific format (FBX, STL, PLY, etc.), use the Format Conversion tool after generation. For example, to convert a Pro edition GLB to FBX, or to extract and re-export a Rapid edition OBJ as GLB, submit the asset URL from the generation result directly to the Convert tool.
# Convert the generated model to FBX
curl -X POST https://api.3daistudio.com/v1/tools/convert/ \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model_url": "https://storage.3daistudio.com/assets/abc123.glb",
"output_format": "fbx"
}'Checking Status
GET/v1/generation-request/<task_id>/status/
Poll this endpoint with the task_id from the submit response. When status is "FINISHED", the results array contains your generated files. Results expire after 24 hours. Note that the asset_type differs by edition: Pro returns "3D_MODEL" (GLB), while Rapid returns "ARCHIVE" (ZIP containing OBJ).
curl https://api.3daistudio.com/v1/generation-request/YOUR_TASK_ID/status/ \
-H "Authorization: Bearer YOUR_API_KEY"// Pro Edition response
{
"status": "FINISHED",
"progress": 100,
"failure_reason": null,
"results": [
{
"asset": "https://storage.3daistudio.com/assets/abc123.glb",
"asset_type": "3D_MODEL",
"metadata": null
}
]
}
// Rapid Edition response
{
"status": "FINISHED",
"progress": 100,
"failure_reason": null,
"results": [
{
"asset": "https://storage.3daistudio.com/assets/def456.zip",
"asset_type": "ARCHIVE",
"metadata": null
}
]
}Credit Costs
| Edition | Base | + PBR | + Multi-View | Max |
|---|---|---|---|---|
| Pro | 60 | +20 | +20 | 100 |
| Rapid | 35 | +20 | N/A | 55 |
Processing Times
The Rapid edition typically takes 2–3 minutes to generate a model. The Pro edition takes longer at 3–6 minutes, depending on the complexity of the input and selected options. Enabling PBR textures or providing multi-view images may add additional processing time.
Errors
Common errors for 3D generation endpoints.
| Status | Error Code | Description |
|---|---|---|
| 402 | insufficient_credits | Not enough credits. Purchase more credits. |
| 429 | rate_limited | Rate limit exceeded. Wait and retry. |
| 400 | validation_failed | Missing or invalid parameters. |