Test-Time Scaling for CAD Generation via Verifier-Free Consensus Selection
2026-08-10 • Computational Engineering, Finance, and Science
Computational Engineering, Finance, and ScienceMachine Learning
AI summaryⓘ
The authors explore how to improve the accuracy of computer programs that create 3D designs from text descriptions. Instead of relying on a separate system to pick the best design from many tries, they propose a method where the chosen design agrees most with the others generated. This approach, called consensus selection, does not need extra training and works with existing tools. Their experiments show that using consensus helps select better designs based on shape and structure compared to random choice or previous selection methods.
Large language modelsParametric CADText-to-CAD generation3D modelingConsensus selectionGeometric agreementTopological agreementChamfer distanceVision-language verifierModel selection
Authors
Aaron Haag, Altay Kaçan, Bertram Fuchs, Oliver Lohse
Abstract
Large language models can write parametric CAD programs from a natural-language description (text-to-CAD generation), but a single sample is often wrong. Increasing test-time compute by sampling multiple candidates only helps if a good candidate can be identified, yet no ground-truth model is available at generation time. Existing systems often require a separate verifier, such as a vision-language judge, to select among candidates. We investigate whether the candidate pool itself provides enough signal for effective selection and a verifier-free alternative. We introduce 3D CAD consensus selection, hereafter consensus selection: sample $N$ parametric CAD programs, compile them to 3D models, and return the candidate that agrees most with the rest of the pool. The method is training-free and compatible with existing CAD agents. We investigate geometric and topological notions of agreement, each of which improves its corresponding evaluation metric. On the exact candidate pools of a state-of-the-art CAD generation method, geometric consensus improves all three geometric metrics over the method's verifier, while topological consensus matches it on topology. Across every tested LLM and prompt variant, geometric consensus also improves geometric accuracy over random selection from the same pool, reducing Chamfer distance by $1-10\%$.