Tencent

GENERATION API CALLS

3D Model Generation

This page is auto-generated from model configurations. Last updated: 2026-09-03.

This reference lists all available Tencent 3d generation models and their parameters. Use these parameter names when calling the Generation API.

Hunyuan 3D 2.1
Hunyuan 3D 3.0 Pro
Hunyuan 3D 3.0 Pro (Multiview)
Hunyuan 3D 3.0 Pro (Sketch)
Hunyuan 3D 3.1 Pro
Hunyuan 3D 3.1 Pro (Multiview)
Hunyuan 3D Part
Hunyuan Polygen 1.5
HY World - Image to Splat
HY World - Multi-view to Splat
HY World - Skybox to Splat
Pixal3D
Tencent Texture Edit
Tencent UV Unwrapping

Hunyuan 3D 2.1

Model ID: model_hunyuan-3d-v2-1

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-3d-v2-1/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`image`	file	Yes	-	-	-	-	Front view to transform in 3D, should remove background to get better results
`paint`	boolean	No	`true`	-	-	-	Paint the mesh with texture.
`steps`	number	No	`30`	10	100	-	The number of steps for the generation
`guidanceScale`	number	No	`5`	1	10	-	Higher values adhere more closely to the input image, while lower values allow for more creative freedom.
`targetFaceNum`	number	No	`40000`	100	1000000	-	The number of faces to target for the generation
`seed`	number	No	-	0	2147483647	-	Use a seed for reproducible results. Leave blank to use a random seed.

Hunyuan 3D 3.0 Pro

Model ID: model_hunyuan-3d-pro-i23d

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-3d-pro-i23d/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`image`	file	Yes	-	-	-	-	Input image. Required if prompt not provided. Size: 128-5000 pixels per side, max 8MB. Formats: jpg, png, jpeg, webp.
`faceCount`	number	No	`500000`	40000	1500000	-	Target number of faces in the generated 3D mesh. Higher values produce finer detail. This parameter is ignored when ‘Optimized Mesh’ mode is enabled.
`generateType`	string	No	`Normal`	-	-	`Normal`, `LowPoly`, `Geometry`	Standard: textured mesh. Optimized Mesh: simplified mesh. Mesh Only: no textures.
`enablePbr`	boolean	No	`false`	-	-	-	Enables Physically Based Rendering (PBR) materials for realistic lighting.
`polygonType`	string	No	`triangle`	-	-	`triangle`, `quadrilateral`	Polygon grid format. Tris for real-time engines; Quads for sculpting.

Hunyuan 3D 3.0 Pro (Multiview)

Model ID: model_hunyuan-3d-pro-multiview

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-3d-pro-multiview/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`frontImage`	file	Yes	-	-	-	-	Front view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`leftImage`	file	No	-	-	-	-	Left view image for multiview generation. Size: 128-5000 pixels per side, max 8MB
`rightImage`	file	No	-	-	-	-	Right view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`backImage`	file	No	-	-	-	-	Back view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`faceCount`	number	No	`500000`	40000	1500000	-	Target number of faces in the generated 3D mesh. Higher values produce finer detail. This parameter is ignored when ‘Optimized Mesh’ mode is enabled.
`generateType`	string	No	`Normal`	-	-	`Normal`, `LowPoly`, `Geometry`	Standard: textured mesh. Optimized Mesh: simplified mesh. Mesh Only: no textures.
`enablePbr`	boolean	No	`false`	-	-	-	Enables Physically Based Rendering (PBR) materials for realistic lighting.
`polygonType`	string	No	`triangle`	-	-	`triangle`, `quadrilateral`	Polygon grid format. Tris for real-time engines; Quads for sculpting.

Hunyuan 3D 3.0 Pro (Sketch)

Model ID: model_hunyuan-3d-pro-sketch

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-3d-pro-sketch/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`prompt`	string	Yes	-	-	-	-	Text prompt for 3D content generation.
`image`	file	Yes	-	-	-	-	Input sketch image. Size: 128-5000 pixels per side, max 8MB.
`faceCount`	number	No	`500000`	40000	1500000	-	Target number of faces in the generated 3D mesh. Higher values produce finer detail.
`enablePbr`	boolean	No	`false`	-	-	-	Enables Physically Based Rendering (PBR) materials for realistic lighting.
`polygonType`	string	No	`triangle`	-	-	`triangle`, `quadrilateral`	Polygon grid format. Tris for real-time engines; Quads for sculpting.

Hunyuan 3D 3.1 Pro

Hunyuan 3D 3.1 (Pro) by Tencent is a state-of-the-art 10B parameter image-to-3D model with 1536³ resolution, hierarchical DiT carving, and enhanced texture color accuracy.

Model ID: model_hunyuan-3d-pro-3-1-i23d

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-3d-pro-3-1-i23d/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`image`	file	Yes	-	-	-	-	Input image. Required if prompt not provided. Size: 128-5000 pixels per side, max 8MB. Formats: jpg, png, jpeg, webp.
`faceCount`	number	No	`500000`	40000	1500000	-	Target number of faces in the generated 3D mesh. Higher values produce finer detail.
`generateType`	string	No	`Normal`	-	-	`Normal`, `Geometry`	Standard: textured mesh. Mesh Only: no textures.
`enablePbr`	boolean	No	`false`	-	-	-	Enables Physically Based Rendering (PBR) materials for realistic lighting.

Hunyuan 3D 3.1 Pro (Multiview)

Hunyuan 3D 3.1 (Multiview) reconciles up to 8 reference images to produce symmetric 3D assets from 120 credits.

Model ID: model_hunyuan-3d-pro-3-1-multiview

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-3d-pro-3-1-multiview/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`frontImage`	file	Yes	-	-	-	-	Front view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`leftImage`	file	No	-	-	-	-	Left view image for multiview generation. Size: 128-5000 pixels per side, max 8MB
`rightImage`	file	No	-	-	-	-	Right view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`backImage`	file	No	-	-	-	-	Back view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`topImage`	file	No	-	-	-	-	Top view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`bottomImage`	file	No	-	-	-	-	Bottom view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`leftFrontImage`	file	No	-	-	-	-	Left front 45° view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`rightFrontImage`	file	No	-	-	-	-	Right front 45° view image for multiview generation. Size: 128-5000 pixels per side, max 8MB.
`faceCount`	number	No	`500000`	40000	1500000	-	Target number of faces in the generated 3D mesh. Higher values produce finer detail.
`generateType`	string	No	`Normal`	-	-	`Normal`, `Geometry`	Standard: textured mesh. Mesh Only: no textures.
`enablePbr`	boolean	No	`false`	-	-	-	Enables Physically Based Rendering (PBR) materials for realistic lighting.

Hunyuan 3D Part

Model ID: model_hunyuan-3d-part

Capabilities: 3d23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-3d-part/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`mesh`	file	Yes	-	-	-	-	Upload a GLB file of the 3D model you want to generate a part of.
`postprocess`	boolean	No	`true`	-	-	-	Post-processingwill merge the small parts according to the threshold. The smaller the threshold, the more parts will be merged.
`postprocessingThreshold`	number	No	`0.95`	0	1	-	Threshold for postprocessing the segmentation mask
`seed`	number	No	-	-	-	-	Use a seed for reproducible results. Leave blank to use a random seed.

Hunyuan Polygen 1.5

Hunyuan Polygen 1.5 by Tencent is an art-grade AI retopology model that converts high-poly 3D meshes into clean, production-ready low-poly assets with quad or triangle topology.

Model ID: model_tencent-smarttopology

Capabilities: 3d23d

LLM Markdown: https://app.scenario.com/api/models/model_tencent-smarttopology/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`file3d`	file	Yes	-	-	-	-	Source 3D file model link. Supported formats: GLB, OBJ. Maximum file size: 200MB.
`polygonType`	string	No	`triangle`	-	-	`triangle`, `quadrilateral`	Polygon type, indicates model surface composition. Triangles: triangle face, Quads: triangular and quadrilateral mixed face
`faceLevel`	string	No	`medium`	-	-	`high`, `medium`, `low`	Reduction level for polygon count. Options: high (minimal reduction), medium (balanced), low (aggressive reduction).

HY World - Image to Splat

Turn a single photo of a place into a navigable 3D Gaussian-splat world. Returns a compact .spz splat for fast preview, with the original .ply available to download.

Model ID: model_hunyuan-world-image-to-splat

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-world-image-to-splat/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`image`	file	Yes	-	-	-	-	A single photo of a PLACE or scene (indoor or outdoor) — NOT a single object. The model expands the surroundings into a full navigable 360° world.
`prompt`	string	No	-	-	-	-	Optional text describing what should fill the unseen surroundings.
`backend`	string	No	`full`	-	-	`full`, `qwen`	Panorama backend. ‘Full’ = HunyuanImage-3 (highest quality, slower & costlier); ‘Qwen’ = faster and cheaper, lower fidelity.
`maxSteps`	number	No	`8000`	50	20000	-	Gaussian-splat training iterations. Higher is marginally sharper; does NOT materially change cost or time.
`applyNavTraj`	boolean	No	`true`	-	-	-	Plan regular navigation fly-through trajectories (better coverage; adds time & cost).
`applyUpRoute`	boolean	No	`true`	-	-	-	Plan wandering ‘up-route’ trajectories (more coverage; adds time & cost).
`applyReconIteration`	boolean	No	`true`	-	-	-	Plan reconstruction-aware iteration passes (higher fidelity; adds time & cost).
`seed`	number	No	-	0	2147483647	-	Seed for reproducible results; leave blank for random.

HY World - Multi-view to Splat

Reconstruct a navigable 3D Gaussian-splat world from multiple photos of the same scene, or a video walkthrough. Returns a compact .spz splat for fast preview, with the original .ply available to download.

Model ID: model_hunyuan-world-multiview-to-splat

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-world-multiview-to-splat/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`images`	file_array	No	-	-	-	-	2–64 overlapping photos of the SAME scene from different viewpoints (no camera setup needed). Provide this OR a video, not both.
`video`	file	No	-	-	-	-	A video walkthrough of the scene; the model samples frames. Provide this OR images, not both.
`targetSize`	number	No	`1036`	256	1920	-	Inference resolution; higher is sharper but slower.
`compressGsMaxPoints`	number	No	`5000000`	10000	20000000	-	Maximum splat points kept after compression; lower = lighter file.

HY World - Skybox to Splat

Turn a 360° skybox (equirectangular panorama) into a navigable 3D Gaussian-splat world. Returns a compact .spz splat for fast preview, with the original .ply available to download.

Model ID: model_hunyuan-world-skybox-to-splat

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_hunyuan-world-skybox-to-splat/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`panorama`	file	Yes	-	-	-	-	A 360° skybox / equirectangular panorama (2:1 aspect ratio) of a scene — e.g. the output of a skybox model.
`maxSteps`	number	No	`8000`	50	20000	-	Gaussian-splat training iterations. Higher is marginally sharper; does NOT materially change cost or time.
`applyNavTraj`	boolean	No	`true`	-	-	-	Plan regular navigation fly-through trajectories (better coverage; adds time & cost).
`applyUpRoute`	boolean	No	`true`	-	-	-	Plan wandering ‘up-route’ trajectories (more coverage; adds time & cost).
`applyReconIteration`	boolean	No	`true`	-	-	-	Plan reconstruction-aware iteration passes (higher fidelity; adds time & cost).
`compressGsMaxPoints`	number	No	`5000000`	10000	20000000	-	Maximum splat points kept after compression; lower = lighter file.

Pixal3D

Pixal3D turns a single image into a high-fidelity 3D model with detailed geometry and realistic textures.

Model ID: model_pixal3d

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_pixal3d/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`image`	file	Yes	-	-	-	-	Input image to convert into a 3D model.
`ssGuidanceStrength`	number	No	`7.5`	0	10	-	How closely the initial 3D structure follows the input image. Higher values are more faithful but may introduce noise.
`ssGuidanceRescale`	number	No	`0.7`	0	1	-	Dampens artifacts from high guidance in the structure stage.
`ssSamplingSteps`	number	No	`36`	1	50	-	Number of denoising steps for initial structure generation.
`shapeSlatGuidanceStrength`	number	No	`7.5`	0	10	-	How closely the detailed geometry follows the input image. Higher values add more detail but may introduce noise.
`shapeSlatGuidanceRescale`	number	No	`0.5`	0	1	-	Dampens artifacts from high guidance in the shape stage.
`shapeSlatSamplingSteps`	number	No	`36`	1	50	-	Number of denoising steps for shape refinement.
`shapeSlatRescaleT`	number	No	`3`	1	6	-	Controls noise schedule sharpness for shape refinement.
`texSlatGuidanceStrength`	number	No	`1`	0	10	-	How closely the texture follows the input image colors. Higher values are more vivid but may oversaturate textures.
`texSlatGuidanceRescale`	number	No	`0`	0	1	-	Dampens artifacts from high guidance in the texture stage.
`texSlatSamplingSteps`	number	No	`36`	1	50	-	Number of denoising steps for texture generation.
`texSlatRescaleT`	number	No	`3`	1	6	-	Controls noise schedule sharpness for texture generation.
`meshScale`	number	No	`1`	0.1	4	-	Scale factor applied to the inferred mesh.
`maxNumTokens`	number	No	`49152`	4096	131072	-	Maximum number of high-resolution tokens used by the shape sampler. Higher values allow finer geometry.
`decimationTarget`	number	No	`200000`	5000	2000000	-	Target number of vertices in the final mesh. Lower values produce smaller files but less detail.
`generationResolution`	number	No	`1536`	-	-	`1024`, `1536`	Internal geometry resolution for generation.
`resolution`	number	No	`4096`	-	-	`1024`, `2048`, `4096`	Texture resolution for the generated 3D model.
`remesh`	boolean	No	`true`	-	-	-	Rebuild the mesh topology for cleaner triangles.
`seed`	number	No	-	-	-	-	Random seed for reproducible results. Leave blank to use a random seed.

Tencent Texture Edit

AI-powered texture editing for 3D models. Apply textures from text prompts or reference images to FBX models. Supports PBR (Physically Based Rendering) when using prompts.

Model ID: model_tencent-texture-edit

Capabilities: 3d23d

LLM Markdown: https://app.scenario.com/api/models/model_tencent-texture-edit/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`file3d`	file	Yes	-	-	-	-	Source 3D model file. Supported format: FBX. Maximum: 100,000 faces.
`image`	file	No	-	-	-	-	Reference image for texture style. Resolution: 128-4096px. Cannot be used with texture prompt.
`prompt`	string	No	-	-	-	-	Texture description prompt. Cannot be used with reference image.

Tencent UV Unwrapping

AI-powered UV unwrapping for 3D models. Generates clean UV maps for FBX, OBJ, and GLB models with up to 30,000 faces.

Model ID: model_tencent-uv-unwrapping

Capabilities: 3d23d

LLM Markdown: https://app.scenario.com/api/models/model_tencent-uv-unwrapping/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
`file3d`	file	Yes	-	-	-	-	Source 3D model file. Maximum: 30,000 faces.