Quantitative Analysis of Technical Debt and Pattern Violation in Large Language Model Architectures

TL;DR

This paper introduces an empirical framework to measure architectural erosion and technical debt in LLM-generated microservices, revealing significant divergence in open-source models' adherence to architectural constraints.

Contribution

It presents the first systematic comparison of state-of-the-art LLMs' ability to produce architecturally compliant microservices using AST parsing and introduces the concept of 'Implementation Laziness.'

Findings

Proprietary models like GPT-5.1 achieve 0% violation rate.

Open-weights models like Llama 3 show 80% architectural violations.

Open-weights models generate 60% fewer logical lines of code.

Abstract

As Large Language Models (LLMs) transition from code completion tools to autonomous system architects, their impact on long-term software maintainability remains unquantified. While existing research benchmarks functional correctness (pass@k), this study presents the first empirical framework to measure "Architectural Erosion" and the accumulation of Technical Debt in AI-synthesized microservices. We conducted a comparative pilot study of three state-of-the-art models (GPT-5.1, Claude 4.5 Sonnet, and Llama 3 8B) by prompting them to implement a standardized Book Lending Microservice under strict Hexagonal Architecture constraints. Utilizing Abstract Syntax Tree (AST) parsing, we find that while proprietary models achieve high architectural conformance (0% violation rate for GPT-5.1), open-weights models exhibit critical divergence. Specifically, Llama 3 demonstrated an 80% Architectural Violation Rate, frequently bypassing interface adapters to create illegal circular dependencies between Domain and Infrastructure layers. Furthermore, we identified a phenomenon of "Implementation Laziness," where open-weights models generated 60% fewer Logical Lines of Code (LLOC) than their proprietary counterparts, effectively omitting complex business logic to satisfy token constraints. These findings suggest that without automated architectural linting, utilizing smaller open-weights models for system scaffolding accelerates the accumulation of structural technical debt.