Vibe Building

2026-10-06

AI Summary

This study addresses the lack of mechanical verification and code compliance in agent-generated building models, which struggle to satisfy physical constraints such as seismic and wind resistance. We propose PE-Loop, a framework that employs a deterministic physics engine as the sole evaluator, mapping building codes into process rewards. By establishing an impartial judge mechanism uninfluenced by the proposer, large language models are restricted to generating discrete revision proposals, enabling closed-loop verification. Simulations on the VB-Bench benchmark demonstrate that PE-Loop achieves the highest verification success rates and holdout-set pass rates for seismic and wind resistance across most backbone models, significantly reducing non-compliant designs. These findings indicate that reliable structural design depends fundamentally on physics engines rather than more capable large language models.

Abstract

Automated building design must comply with seismic and wind codes and satisfy structural mechanics constraints, yet most existing agents produce visually plausible models without verification grounded in mechanical analysis and code compliance. We introduce the Vibe Building task and propose PE-Loop (Physics-Engine-in-the-Loop), an agent in which a deterministic physics engine is the sole source of evaluation signals, mapping code constraints to a physics process reward, while the language model is confined to proposing discrete revisions (section menu, topology, and lateral system). Designs are verified by held-out seismic and wind time-history checks and a constructability gate. On VB-Bench, 3,577 physics-adjudicated building instances across six code families, PE-Loop achieves the highest verified success rate under three of four backbone LLMs, the highest held-out seismic pass rate under all four, and the highest held-out wind pass rate under three. Replacing the physics verdict with a language-model judge, all else fixed, leaves 58.43% of accepted designs noncompliant. These results suggest that reliable structural design rests less on a stronger LLM proposer than on an adjudicator the proposer cannot influence, a division of labor for agents whose outputs must hold up in the physical world.

Problem

• automated building design
• structural mechanics constraints
• seismic and wind code compliance
• physics verification

Innovation

• Physics-Engine-in-the-Loop
• Automated building design
• Code compliance verification
• Process reward
• Agent adjudicator