Arbor
A solid pick for self-hosting and full data control.
Overview
Arbor is Stability AI's answer (with the University of Tübingen) to a well-known limit of generative 3D: text-to-3D models produce convincing objects but offer no control over the space an object should occupy or avoid. Released in late June 2026 as an open-source release, Arbor introduces constraint meshes as a native control interface: "hull" regions where geometry must exist, avoidance regions that must stay empty, and touch regions the object must contact. Concretely: a chair that fits a given seating envelope, a mechanical part exposing a precise contact surface, a game prop leaving the clearance a player needs. These are spatial intents known before generation starts — intents neither a prompt nor a reference image can carry. Technically, Arbor is a trainable attachment grafted onto TRELLIS.2 (whose generator stays frozen), converting constraint meshes into tokens via the O-Voxel encoders. In a user study (27 participants, 404 trials), Arbor variants won 59.2% of pairwise choices against competing methods. The release is inference-only: public pipeline, model loading, mesh export and a Blender add-on, under the Stability AI Community License. It requires Linux and an NVIDIA GPU. It's still more research tool than product, but it prefigures the next step in generative 3D: generation under geometric specification.
Skill profile
Not disclosed
Strengths
- Has a free tier
- Open-source and self-hostable
Who is it for
- you want to control cost or self-host
- your data must stay in the EU
Ideal use cases
- 3D assets under spatial constraints (envelopes, clearances)
- Product and furniture design with imposed contact surfaces
- Level design: objects that respect the play space
Access & availability
Key specifications
What it costs per month
Estimate from sourced API prices and a hardware tier for self-hosting. Assumptions are editable.
Assumptions: 36-month amortisation, 0.25 €/kWh, 22 days/month, 1 $ = 0.92 €. Excludes labour, software licences, redundancy.
nAIvigate Trust Index
What the vendor offers and documents for professional use in Europe. An undocumented criterion counts as zero and stays visible: the score rises when the vendor publishes, not when we assume.
19/19 criteria documented · 100 % · 12 of them set by rule
See the criteria▾
Trust Index badge
Show this score in your docs, README or security page. It updates automatically.
[](https://naivigate.com/en/model/arbor)Data & compliance
What we know, line by line, with source and date. These lines feed the Data and Enterprise families of the Trust Index above. GDPR compliance depends on your processing, not on the model alone.
| EU data residency | Yes |
| Data processing agreement (DPA) | N/A· self-hosted: data never leaves |
| Transfer outside the EU | No |
| Training on your data | No |
| Configurable retention | N/A· self-hosted: data never leaves |
| Zero Data Retention | N/A· self-hosted: data never leaves |
| Self-hostable | Yes |
| Declared hosting | self |
Quality of this card’s data
This score measures the quality of the data on this card: price, context, Elo, licence, documented, sourced and fresh. It measures neither the model nor its compliance: for that, see the Trust Index above.
3/8 documented · 0/8 primary sources · 2 older than 90 days
| API pricingopen weights, no API pricing | missing | |
| Context window | nAIvigate assessment | |
| Arena Elo | missing | |
| Benchmarks (MMLU, GPQA, SWE-bench…) | missing | |
| Licence | nAIvigate assessment | |
| Data & complianceself-hosted: derived | derived | |
| Adoption (Hugging Face) | missing | |
| Hardware requirements (quantizations)computed from model size | missing |
Advanced data · for expertsArchitecture, modalities, detailed cost, full benchmarks▾
| Arena Elo | — |
| MMLU | — |
| GPQA | — |
| HumanEval | — |
| SWE-Bench | — |
| MATH | — |