一次性积分包 $14.90 起

Open-Source Image-to-3D: TRELLIS.2 vs Hunyuan3D, TripoSR

2026/01/27
阅读约 7 分钟

试试 TRELLIS.2 AI 3D 生成器

把图片快速变成可下载的 3D 模型,适合产品图、游戏资产和原型验证。

立即生成 3D 模型

What This Review Covers

This page reviews open-weight image-to-3D releases: models whose weights you can download and run on your own hardware. It is not a review of hosted services — for those, use the workflow comparison instead. If you only want to know what Microsoft's release is and how this site relates to it, the Microsoft TRELLIS.2 explainer answers that directly.

Every product fact below comes from the project's own repository or documentation, linked inline. This page does not publish an unrecorded benchmark, does not rank the projects by output quality, and does not borrow numbers from a different release.

The Four Releases Side By Side

ProjectLicence published in the repositoryWhat it ships
TRELLIS.2 (Microsoft)MIT4B model weights, inference and training code
Hunyuan3D 2.1 (Tencent)Tencent Hunyuan 3D 2.1 Community License3.3B shape + 2B texture weights, training code
TripoSR (Tripo AI + Stability AI)MITSource code, pretrained model, demo script
Stable Fast 3D (Stability AI)Stability AI Community LicenseWeights on a gated Hugging Face repo, code, ComfyUI nodes

TRELLIS.2: The Microsoft Release

Microsoft's official repository describes TRELLIS.2 as a 4-billion-parameter image-to-3D model built on a field-free sparse voxel structure it calls O-Voxel, which the project credits with handling open surfaces (clothing, leaves), non-manifold geometry, and internal enclosed structures without a lossy conversion step. Published output resolutions are 512³, 1024³, and 1536³, with full PBR materials — base color, roughness, metallic, and opacity — and the repository is MIT-licensed.

Setup is the demanding part. The project states the code is currently tested only on Linux, needs an NVIDIA GPU with at least 24 GB of memory, and has been verified on A100 and H100 cards. The documented environment is Python 3.8 or later with PyTorch 2.6.0 on CUDA 12.4, flash-attn as the default attention backend, and xformers as the fallback for GPUs that do not support it. Before installing, read the current README rather than an older tutorial: the earlier Microsoft TRELLIS project is a different release, and its model names, flags, and output modes do not transfer.

The pretrained checkpoint is published as microsoft/TRELLIS.2-4B on Hugging Face with a demo Space, and Microsoft's own inference timings (about 3 s / 17 s / 60 s at the three resolutions) are measured on an H100, so treat them as model-runtime figures rather than end-to-end production latency. Export goes through the project's O-Voxel post-processing into GLB, with parameters for decimation target and texture size; note that exported GLBs are opaque by default and the alpha channel must be connected manually if you need transparency.

Hunyuan3D 2.1: Weights, Training Code, and a Territorial Licence

Tencent's Hunyuan3D 2.1 repository publishes two components: Hunyuan3D-Shape-v2-1 (3.3B parameters, image to shape) and Hunyuan3D-Paint-v2-1 (2B parameters, texture synthesis). The project describes it as the first release of its line to ship full model weights and training code, and its texture pipeline models PBR materials rather than only RGB colour.

The README states memory needs of about 10 GB VRAM for shape generation, 21 GB for texture generation, and 29 GB for both, and documents macOS, Windows, and Linux support with Python 3.10 and PyTorch 2.5.1. Tencent also operates a hosted workspace at 3d.hunyuan.tencent.com, so the same brand appears on both the open-weight and hosted sides of a shortlist.

The licence is the constraint to read before planning a deployment. The Tencent Hunyuan 3D 2.1 Community License excludes the European Union, the United Kingdom, and South Korea from its licensed territory; products above 1 million monthly active users need a separate licence from Tencent; and anyone offering the model as a service must disclose the real provider and state that Tencent is not affiliated with or endorsing it.

TripoSR: The Lightweight Baseline

TripoSR is the official codebase for a feedforward single-image reconstruction model developed by Tripo AI and Stability AI. The repository ships source code, pretrained weights, and an interactive demo under the MIT licence, and documents a default footprint of about 6 GB VRAM for a single input image — the lightest of the four to stand up. The project states that it can reconstruct a model in under 0.5 seconds on an A100; that is the vendor's own published figure on named hardware, not an independent measurement, and it is not the time you will spend getting a usable asset into a destination tool.

Output is a mesh with vertex colours by default, with an optional baked texture pass and a configurable texture resolution. The repository does not document a remeshing step, so if topology matters to your pipeline, plan to retopologise separately.

Stable Fast 3D: TripoSR With UV Unwrapping and Lighting Estimates

Stable Fast 3D (SF3D) is Stability AI's successor in the same family: the README states it is based on TripoSR but explicitly optimised to produce meshes without reconstruction artefacts alongside textures with UV unwrapping, and it predicts material parameters so assets can be dropped into a renderer. It exports GLB, uses roughly the same 6 GB VRAM footprint, and exposes a texture resolution plus remesh options (none, triangle, or quad) with a target vertex count. The repository also ships official ComfyUI nodes.

Two practical differences matter. First, the weights sit behind a gated Hugging Face repository, so access must be requested and a token configured before the first run. Second, the licence is the Stability AI Community License: free for research and non-commercial use, and free for commercial use while the organisation stays below 1,000,000 USD in annual revenue, with registration required for commercial use and a separate enterprise licence above that threshold. The README marks Mac (MPS) and Windows support as experimental, with a CPU fallback.

Hardware Is The Real Filter

Sort the four by what you can actually run:

  • Linux, 24 GB+ NVIDIA GPU: TRELLIS.2, the only one of the four whose repository publishes explicit 512³–1536³ output modes.
  • Up to 29 GB VRAM on macOS, Windows, or Linux: Hunyuan3D 2.1, with a two-stage shape-plus-texture pipeline.
  • About 6 GB VRAM: TripoSR, and Stable Fast 3D, which additionally documents experimental Mac and Windows support and a CPU fallback.

Add the licence to the hardware column: an MIT or gated-community model is not interchangeable with a licence that excludes your territory or charges for scale.

What None Of Them Will Do For You

A single image does not contain the back, the underside, the true dimensions, or the internal structure of an object. Every model here estimates those regions, so inspect the result from several angles before trusting it. None of the four certifies an asset for a specific destination either: topology, UVs, normals, scale, material channels, and wall thickness still need checking in the tool where the asset will live, and a mesh that looks correct in a viewer can still fail a slicer or an engine import.

Generated assets are drafts. Budget for retopology, UV repair, or scaling work, and never use a single-image reconstruction as dimensional evidence for a functional part.

How To Choose

Run the same small test on the model your hardware and licence allow, using your own images, and record failures as well as successes. The comparison protocol covers the accepted-asset calculation and the fields to capture per job.

If you do not want to run any of this locally, the hosted TRELLIS.2 workflow on this site takes one image, tracks the provider task, and returns a GLB — start with a 512 job and inspect the file before spending more credits.

Read the official Microsoft TRELLIS.2 repository

Test the independent hosted GLB workflow

TRELLIS.2 Editorial Team

Meshy Hermes Agent Plugin vs Trellis Mini $4.99 Inspect-First GLB

2026/10/31

Meshy Hermes Agent Plugin vs Trellis Mini $4.99 Inspect-First GLB

Meshy Agent/plugin subscription path vs Trellis Mini $4.99/150cr for one 512³ inspect-first GLB. Independent host, not Microsoft. Verified 2026-10-31.

Meshy First-Month ~50% Off vs Trellis Mini $4.99 GLB

2026/10/29

Meshy First-Month ~50% Off vs Trellis Mini $4.99 GLB

Meshy Pro shows ~$10 first month (~50% off). Trellis Mini is $4.99/150cr = one 512³ inspect-first GLB. Independent host, not Microsoft. Confirm checkout.

Print-Ready from Chat (Meshy Agent) vs Inspect-First $4.99 GLB

2026/10/27

Print-Ready from Chat (Meshy Agent) vs Inspect-First $4.99 GLB

Meshy Agent print-ready chat vs inspect-first Mini $4.99/150cr = one 512³ GLB on trellis-2.com. Independent host, not Microsoft. Checklist before slicer.

When You Don’t Need a 3D Agent: First Accepted GLB Checklist

2026/10/27

When You Don’t Need a 3D Agent: First Accepted GLB Checklist

Skip Meshy Agent chat when one photo → one inspectable GLB. Mini $4.99/150cr = one 512³ on trellis-2.com. Independent host, not Microsoft. Checklist first.

Hosted TRELLIS.2 vs Meshy Subscription: Cost Per Accepted GLB

2026/10/26

Hosted TRELLIS.2 vs Meshy Subscription: Cost Per Accepted GLB

One-time never-expire credits on trellis-2.com (150/300/500 per resolution) vs Meshy monthly pools. Compare cost per accepted GLB — independent host, not Microsoft.

📝

2026/10/08

2026 年免费在线 3D 查看器对比:GLB、STL、OBJ、FBX

六种几乎不用安装就能打开的 3D 文件方案,按格式、是否上传、是否要登录和限制对比,包含 TRELLIS.2 的浏览器查看器。

📝

2026/10/04

打印前怎么检查 3D 网格:在浏览器里先做一遍

在把生成的网格转成 STL、送进切片软件之前,先在浏览器里检查一遍:提前发现破洞、翻转的法线和过薄的壁。

📝

2026/10/04

在浏览器里把 GLB 草模转成 STL:会留下什么,会丢掉什么

在浏览器里把 GLB 草模转成 STL,弄清哪些数据留下、哪些数据丢掉,并在切片之前先确认三角面片,避免薄壁在打印时才暴露问题。

📝

2026/10/04

如何在线打开 3MF 文件:先看清包里到底是什么形状

在浏览器里打开 3MF 包,旋转网格,并分清这个文件带有什么、不带有什么,再决定要不要进一个单独的切片步骤。

📝

2026/10/04

如何在线打开 FBX 文件:形体能看,场景不保证

把 FBX 拖进浏览器,旋转查看网格,同时明确这个预览会显示哪些场景数据、不会显示哪些,再决定是否要在别处转换或编辑。

📝

2026/10/04

如何在线打开 GLB 文件:带材质的浏览器预览

在线打开 GLB 文件,包括本站从一张图生成的带纹理草模,并学会在转换或打印之前判断哪些面是模型估算出来的。

📝

2026/10/04

如何在线打开 glTF 文件:先确认 .bin 和贴图有没有一起到

在线打开 glTF 文件,并弄清楚为什么单独的 JSON 会打不开:缓冲区或贴图被落在别处时,先把交付包补全再预览。

📝

2026/10/04

如何在线打开 OBJ 文件:先看网格,再看材质

在浏览器里打开 OBJ 文件,看清楚的只有你拖进来的那个网格;同时知道什么时候该保留 GLB 的贴图、什么时候才需要转成 OBJ。

📝

2026/10/04

如何在线打开 PLY 文件:网格、扫描与点云的区别

在浏览器里打开 PLY 网格或扫描件,顺带看清顶点颜色能显示到什么程度,以及一个只有点、没有面的文件为什么画不成实体。

📝

2026/10/04

如何在线打开 STL 文件:在浏览器里查看三角网格

把 STL 拖进浏览器就能旋转查看三角面片,同时看清为什么预览是灰色、没有打印参数,以及切片前必须自己确认的尺寸和壁厚。

How to Turn a Photo into a 3D Model

2026/10/04

How to Turn a Photo into a 3D Model

Turn one photo into a downloadable GLB draft, then inspect the estimated sides before you convert or print so thin walls and holes show up early on.

Open-Source 3D AI in 2026: A Verification-First Guide

2026/01/27

Open-Source 3D AI in 2026: A Verification-First Guide

Compare hosted trials and open-source image-to-3D models without invented rankings, then check current TRELLIS.2 facts and a cost-per-accepted-asset test.

ComfyUI and TRELLIS.2: A Safe Local Setup Checklist (2026)

2026/01/27

ComfyUI and TRELLIS.2: A Safe Local Setup Checklist (2026)

Official TRELLIS.2 requirements, ComfyUI wrapper limits, install safety, model checks, workflow layout and troubleshooting for local image-to-3D.

Open-Source Image-to-3D: TRELLIS.2 vs Hunyuan3D, TripoSR

2026/01/27

Open-Source Image-to-3D: TRELLIS.2 vs Hunyuan3D, TripoSR

Source-based review of open-weight image-to-3D releases you can run yourself: TRELLIS.2, Hunyuan3D 2.1, TripoSR and Stable Fast 3D, licences included.

TRELLIS.2 to Blender: Import GLB, Inspect, Clean, and Export

2026/01/27

TRELLIS.2 to Blender: Import GLB, Inspect, Clean, and Export

A practical TRELLIS.2-to-Blender workflow for importing the hosted GLB, checking transforms, geometry and PBR materials, then exporting the scene.

TRELLIS.2 Tutorial: From One Image to a Reviewed GLB

2026/01/27

TRELLIS.2 Tutorial: From One Image to a Reviewed GLB

Step-by-step TRELLIS.2 guide: input selection, 512/1024/1536 credit cost, tracked generation, GLB inspection, failure refunds and destination checks.

Can TRELLIS.2 Models Be 3D Printed? Mesh Validation

2026/01/27

Can TRELLIS.2 Models Be 3D Printed? Mesh Validation

A practical TRELLIS.2 print-readiness workflow: manifoldness, holes, normals, wall thickness, scale, STL conversion, slicing and small test prints.

TRELLIS.2 in a VFX Pipeline: An Evaluation Checklist

2026/01/27

TRELLIS.2 in a VFX Pipeline: An Evaluation Checklist

A transparent workflow for evaluating TRELLIS.2 output for VFX, previs, set dressing, and production handoff, without any fabricated artist testimonials.

Trellis 2 Quickstart: Your First Image-to-3D Draft

2026/01/27

Trellis 2 Quickstart: Your First Image-to-3D Draft

A truthful five-minute TRELLIS.2 quickstart: sign in, buy credits, start a 512 job, follow live job status, inspect the result, and download a GLB draft.

📝

2025/12/18

TRELLIS.2(trellis-2)完整指南:图像转3D、PBR材质、GLB导出

一篇讲人话的 trellis-2 / trellis 2 / trellis.2 教程:TRELLIS.2 是什么、怎么把图片变3D、哪些图最适合、核心参数怎么选、以及GLB导出后怎么进Blender/Unity。

← Swipe →
Open-Source Image-to-3D: TRELLIS.2 vs Hunyuan3D, TripoSR