DeepEye-SQL: A Software-Engineering-Inspired Text-to-SQL Framework

TL;DR

DeepEye-SQL introduces a software-engineering-inspired framework for Text-to-SQL tasks, emphasizing structured, verifiable orchestration over pure language generation, leading to improved accuracy without model fine-tuning.

Contribution

It redefines Text-to-SQL as a software development process with stages like schema linking, fault tolerance, verification, and confidence-based selection, outperforming larger models.

Findings

Achieves 89.8% accuracy on Spider-Test.

Outperforms state-of-the-art solutions without fine-tuning.

Demonstrates the importance of structured orchestration over model scaling.

Abstract

Large language models (LLMs) have advanced Text-to-SQL, yet existing solutions still fall short of system-level reliability. The limitation is not merely in individual modules -- e.g., schema linking, reasoning, and verification -- but more critically in the lack of structured orchestration that enforces correctness across the entire workflow. This gap motivates a paradigm shift: treating Text-to-SQL not as free-form language generation but as a software-engineering problem that demands structured, verifiable orchestration. We present DeepEye-SQL, a software-engineering-inspired framework that reframes Text-to-SQL as the development of a small software program, executed through a verifiable process guided by the Software Development Life Cycle (SDLC). DeepEye-SQL integrates four synergistic stages: it grounds user intent through robust schema linking, enforcing relational closure; enhances fault tolerance with N-version SQL generation; ensures deterministic verification via a ``Syntax-Logic-Quality'' tool-chain that intercepts errors pre-execution; and introduces confidence-aware selection that leverages execution-guided adjudication to resolve ambiguity beyond simple majority voting. Leveraging open-source MoE LLMs (~30B total, ~3B activated parameters) without any fine-tuning, DeepEye-SQL achieves 73.5% execution accuracy on BIRD-Dev, 75.07% on the official BIRD-Test leaderboard, and 89.8% on Spider-Test, outperforming state-of-the-art solutions that rely on larger models or extensive training. This highlights that principled orchestration, rather than LLM scaling alone, is key to achieving system-level reliability in Text-to-SQL.