LLM NL2SQL Robustness: Surface Noise vs. Linguistic Variation in Traditional and Agentic Settings

TL;DR

This paper evaluates the robustness of large language models in translating natural language to SQL under various noisy and variable conditions, revealing strengths and vulnerabilities in different settings.

Contribution

Introduces a comprehensive robustness benchmark for NL2SQL systems and compares multiple state-of-the-art LLMs across traditional and agentic scenarios.

Findings

Models are robust against some perturbations but struggle with surface noise and linguistic variation.

Surface noise impacts traditional pipelines more significantly.

Linguistic variation challenges agentic NL2SQL systems.

Abstract

Robustness evaluation for Natural Language to SQL (NL2SQL) systems is essential because real-world database environments are dynamic, noisy, and continuously evolving, whereas conventional benchmark evaluations typically assume static schemas and well-formed user inputs. In this work, we introduce a robustness evaluation benchmark containing approximately ten types of perturbations and conduct evaluations under both traditional and agentic settings. We assess multiple state-of-the-art large language models (LLMs), including Grok-4.1, Gemini-3-Pro, Claude-Opus-4.6, and GPT-5.2. Our results show that these models generally maintain strong performance under several perturbations; however, notable performance degradation is observed for surface-level noise (e.g., character-level corruption) and linguistic variation that preserves semantics while altering lexical or syntactic forms. Furthermore, we observe that surface-level noise causes larger performance drops in traditional pipelines, whereas linguistic variation presents greater challenges in agentic settings. These findings highlight the remaining challenges in achieving robust NL2SQL systems, particularly in handling linguistic variability.