Metraipolis › archive › dc_e0c3ded7d0df
ReToolSQL: Agentic Reinforcement Learning for Robust Text-to-SQL — arXiv:2608.27796
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.27796
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T10:47:11.775Z
- SHA-256 of the archived text
399856f3e357c2d3f74a2a0eaca3c3cbe9c7a4ba4f7ef2610db231eb9eee15bc- Stable URL of this page
https://metraipolis.com/documento/dc_e0c3ded7d0df
How to cite
ReToolSQL: Agentic Reinforcement Learning for Robust Text-to-SQL — arXiv:2608.27796 — arXiv (en), https://arxiv.org/abs/2608.27796 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e0c3ded7d0df (Metraipolis archive dc_e0c3ded7d0df, sha256 399856f3e357c2d3f74a2a0eaca3c3cbe9c7a4ba4f7ef2610db231eb9eee15bc)
BibTeX
@misc{dce0c3ded7d0df,
title = {ReToolSQL: Agentic Reinforcement Learning for Robust Text-to-SQL — arXiv:2608.27796},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.27796},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e0c3ded7d0df; sha256 399856f3e357c2d3f74a2a0eaca3c3cbe9c7a4ba4f7ef2610db231eb9eee15bc},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.27796. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Recent work has shown that reinforcement learning from execution feedback can substantially improve text-to-SQL performance, often enabling smaller models to match or exceed much larger systems. However, most existing approaches treat SQL generation as a single-turn task, limiting the model's ability to recover from errors through iterative refinement. We present ReToolSQL, a two-stage training framework for text-to-SQL that combines (i) a supervised warm-start on rejection-sampled reasoning traces with (ii) agentic reinforcement fine-tuning (RFT) over multi-turn tool-use trajectories. The key insight is that the two stages act on complementary axes, the supervised fine-tuning (SFT) on verified privileged-teacher traces expands the set of solvable questions (raising pass@k coverage on the hardest cases), while RFT converts that expanded capability into higher single-pass accuracy by teaching the model when to verify, what evidence to retrieve, and how to repair faulty SQL from execution feedback. Applied to Gemma 4 instruction-tuned (31B), RFT alone achieves 73.66% execution accuracy (EX) on the BIRD-SQL development benchmark (74.12% EX with self-consistency). Initializing RFT from the SFT checkpoint (SFT$\to$RFT) yields our strongest model at 74.32% EX single-pass and 74.77% EX with self-consistency. At the time of writing, this ranked first on the BIRD single-model development-set leaderboard. The approach uses composite rewards anchored on execution correctness, requires no human annotation beyond the benchmark itself, and operates within a single dense 31B model, showing that a properly designed SFT$\to$RFT pipeline over tool-use trajectories is a practical path toward robust enterprise-grade text-to-SQL. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-31 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.27796 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T10:33:42.348Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-31?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).