# Human-Level Text-to-SQL via Reinforcement Learning on Verified Data, Without Pipeline Engineering

arXiv: https://arxiv.org/abs/2603.20004 · HTML: https://mlsanity.com/paper/2603.20004/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in RL training of language and agent models, one of 15 directions in Deep reinforcement learning methods.
- Builds on 31 of its 103 references that are in the map.
- First posted on arXiv 20 Mar 2026.

## Placed in

- Deep reinforcement learning methods / [RL training of language and agent models](https://mlsanity.com/deeprl/rl-training-of-language-and-agent-models/)

## Builds on (most cited first, 30 of 31)

- [Can LLM Already Serve as A Database Interface? A BIg Bench for Large-Scale Database Grounded Text-to-SQLs](https://mlsanity.com/paper/2305.03111/) (2023)
- [Why Do Multi-Agent LLM Systems Fail?](https://mlsanity.com/paper/2503.13657/) (2025)
- [Text-to-SQL Empowered by Large Language Models: A Benchmark Evaluation](https://mlsanity.com/paper/2308.15363/) (2023)
- [MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention](https://mlsanity.com/paper/2506.13585/) (2025)
- [Spider 2.0: Evaluating Language Models on Real-World Enterprise Text-to-SQL Workflows](https://mlsanity.com/paper/2411.07763/) (2024)
- [CodeS: Towards Building Open-source Language Models for Text-to-SQL](https://mlsanity.com/paper/2402.16347/) (2024)
- [CHESS: Contextual Harnessing for Efficient SQL Synthesis](https://mlsanity.com/paper/2405.16755/) (2024)
- [C3: Zero-shot Text-to-SQL with ChatGPT](https://mlsanity.com/paper/2307.07306/) (2023)
- [CHASE-SQL: Multi-Path Reasoning and Preference Optimized Candidate Selection in Text-to-SQL](https://mlsanity.com/paper/2410.01943/) (2024)
- [The Death of Schema Linking? Text-to-SQL in the Age of Well-Reasoned Language Models](https://mlsanity.com/paper/2408.07702/) (2024)
- [Reasoning-SQL: Reinforcement Learning with SQL Tailored Partial Rewards for Reasoning-Enhanced Text-to-SQL](https://mlsanity.com/paper/2503.23157/) (2025)
- [OpenSearch-SQL: Enhancing Text-to-SQL with Dynamic Few-shot and Consistency Alignment](https://mlsanity.com/paper/2502.14913/) (2025)
- [XiYan-SQL: A Novel Multi-Generator Framework For Text-to-SQL](https://mlsanity.com/paper/2507.04701/) (2025)
- [Arctic-Text2SQL-R1: Simple Rewards, Strong Reasoning in Text-to-SQL](https://mlsanity.com/paper/2505.20315/) (2025)
- [ReFoRCE: A Text-to-SQL Agent with Self-Refinement, Consensus Enforcement, and Column Exploration](https://mlsanity.com/paper/2502.00675/) (2025)
- [Understanding the Effects of Noise in Text-to-SQL: An Examination of the BIRD-Bench Benchmark](https://mlsanity.com/paper/2402.12243/) (2024)
- [Alpha-SQL: Zero-Shot Text-to-SQL using Monte Carlo Tree Search](https://mlsanity.com/paper/2502.17248/) (2025)
- [A Comparative Survey of Recent Natural Language Interfaces for Databases](https://mlsanity.com/paper/1906.08990/) (2019)
- [NL2SQL-BUGs: A Benchmark for Detecting Semantic Errors in NL2SQL Translation](https://mlsanity.com/paper/2503.11984/) (2025)
- [Evaluating Cross-Domain Text-to-SQL Models and Benchmarks](https://mlsanity.com/paper/2310.18538/) (2023)
- [Automatic Metadata Extraction for Text-to-SQL](https://mlsanity.com/paper/2505.19988/) (2025)
- [ExCoT: Optimizing Reasoning for Text-to-SQL with Execution Feedback](https://mlsanity.com/paper/2503.19988/) (2025)
- [Agentar-Scale-SQL: Advancing Text-to-SQL through Orchestrated Test-Time Scaling](https://mlsanity.com/paper/2509.24403/) (2025)
- [Reliable Text-to-SQL with Adaptive Abstention](https://mlsanity.com/paper/2501.10858/) (2025)
- [SHARE: An SLM-based Hierarchical Action CorREction Assistant for Text-to-SQL](https://mlsanity.com/paper/2506.00391/) (2025)
- [Cheaper, Better, Faster, Stronger: Robust Text-to-SQL without Chain-of-Thought or Fine-Tuning](https://mlsanity.com/paper/2505.14174/) (2025)
- [Pervasive Annotation Errors Break Text-to-SQL Benchmarks and Leaderboards](https://mlsanity.com/paper/2601.08778/) (2026)
- [SiriusBI: A Comprehensive LLM-Powered Solution for Data Analytics in Business Intelligence](https://mlsanity.com/paper/2411.06102/) (2024)
- [SLM-SQL: An Exploration of Small Language Models for Text-to-SQL](https://mlsanity.com/paper/2507.22478/) (2025)
- [VeriEQL: Bounded Equivalence Verification for Complex SQL Queries with Integrity Constraints](https://mlsanity.com/paper/2403.03193/) (2024)

## Built on by (most cited first, 1 of 1)

- [STRATOS: Bridging the Symbolic-to-Numeric Gap in Spatio-Temporal Text-to-SQL for Meteorological Data](https://mlsanity.com/paper/2607.03501/) (2026)

## Abstract (arXiv)

Translating natural language questions to SQL queries (Text-to-SQL) is a long-standing problem in database research. Recent efforts have focused on improving accuracy by building increasingly complex multi-stage large LLM pipelines, layering task decomposition, schema linking, and model-based query selection on top of an LLM. Despite this growing complexity, a substantial gap (>10%) between such systems and human experts persists on benchmarks, suggesting that pipeline engineering alone has hit a ceiling. We show that human-level Text-to-SQL performance is achievable by fine-tuning an LLM using RLVR on clean data, without pipeline components. In this paper, we identified the dominant bottleneck for RLVR on Text-to-SQL: existing training data contains pervasive annotation errors that mislead optimization. To address this, we developed a multi-round, expert-driven verification pipeline and used it to curate BIRD-Platinum, a dataset of 2.5k verified instances sampled from BIRD Train, with errors corrected in 61% of instances. We show that fine-tuning Qwen3-235B on BIRD-Platinum yields consistent improvements (11-16%) over BIRD Train on Arcwise-Plat (an expert-verified version of BIRD) and Spider2, outperforming SOTA open-source systems by 0.6-16%. Furthermore, we diagnosed two failure modes of standard RLVR on Text-to-SQL. We find that (1) result-based rewards have non-trivial false positive rates, and (2) models systematically ignore the external knowledge in BIRD-style problems. To address them, we propose ReViSQL-BIRD, a specialized reward shaping method that combines result-based rewards with SQL equivalence verification and leverages process rewards for incentivizing external-knowledge use. We fine-tuned Kimi-K2.6 with ReViSQL-BIRD. On Arcwise-Plat, ReViSQL-BIRD-K2.6 is the first method to achieve human-level accuracy (92.96%), outperforming top five open-source systems by 10-22%.