Academia

This page is auto-generated from model configurations. Last updated: 2026-04-09.

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

• PartCrafter
• ReconViaGen 0.5
• Ultrashape 1.0

---

PartCrafter

PartCrafter is a 3D model generator that creates 3D models from images.

Model ID: model_partcrafter

Capabilities: img23d

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

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
images	file	Yes	-	-	-	-	Input image to generate 3D asset from
numParts	number	No	6	1	16	-	Number of parts to generate
steps	number	No	75	1	100	-	Number of inference steps.
guidanceScale	number	No	7	1	10	-	Higher values will keep output closer to the input image
numTokens	number	No	4096	1024	4096	-	Number of tokens
seed	number	No	-	0	2147483647	-	Use a seed for reproducible results. Leave blank to use a random seed.

ReconViaGen 0.5

Generate a textured 3D model from one or more images using ReconViaGen 0.5.

Model ID: model_reconviagen-0-5

Capabilities: img23d

LLM Markdown: https://app.scenario.com/api/models/model_reconviagen-0-5/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
images	file_array	Yes	-	-	-	-	One or more views of the same object for 3D reconstruction.
resolution	number	No	1024	-	-	512, 1024, 1536	Output resolution. Higher values produce more detail.
multiImageStrategy	string	No	adaptive_guidance_weight	-	-	average_right, weighted_average, sequential, average, adaptive_guidance_weight, fixed_guidance_rescale	How multiple input views are combined.
ssSource	string	No	mesh	-	-	direct, mesh, mvtrellis2	Source strategy for sparse structure generation.
textureSize	number	No	2048	-	-	1024, 2048, 4096	Resolution of the baked texture map.
decimationTarget	number	No	500000	5000	2000000	-	Target vertex count for the final mesh.
seed	number	No	-	0	2147483647	-	Random seed for reproducibility.

Ultrashape 1.0

UltraShape-1.0 is a 3D diffusion framework that generates high-fidelity 3D geometry through coarse-to-fine geometric refinement.

Model ID: model_ultrashape-1-0

Capabilities: 3d23d

LLM Markdown: https://app.scenario.com/api/models/model_ultrashape-1-0/markdown

Parameter	Type	Required	Default	Min	Max	Allowed Values	Description
image	file	Yes	-	-	-	-	Reference image for mesh refinement.
model	file	Yes	-	-	-	-	The coarse mesh to refine
numInferenceSteps	number	No	50	1	50	-	Number of inference steps.
octreeResolution	number	No	1024	128	1024	-	Marching cubes resolution.
seed	number	No	-	0	-	-	Seed value for randomization, leave blank to use a random seed