MTSQL-R1: Towards Long-Horizon Multi-Turn Text-to-SQL via Agentic Training
About
Multi-turn Text-to-SQL aims to translate a user's conversational utterances into executable SQL while preserving dialogue coherence and grounding to the target schema. However, most existing systems only regard this task as a simple text translation task and follow a short-horizon paradigm, generating a query per turn without execution, explicit verification, and refinement, which leads to non-executable or incoherent outputs. We present MTSQL-R1, an agentic training framework for long-horizon multi-turn Text-to-SQL. We cast the task as a Markov Decision Process (MDP) in which an agent interacts with (i) a database for execution feedback and (ii) a persistent dialogue memory for coherence verification, performing an iterative propose to execute -> verify -> refine cycle until all checks pass. Experiments on COSQL and SPARC demonstrate that MTSQL-R1 consistently outperforms strong baselines, highlighting the importance of environment-driven verification and memory-guided refinement for conversational semantic parsing. Full recipes (including code, trained models, logs, reasoning trajectories, etc.) will be released after the internal review to contribute to community research.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Text-to-SQL | Spider | Exec Acc (All)84.2 | 139 | |
| Multi-turn Text-to-SQL | CoSQL In-domain | Exact Match65.2 | 29 | |
| Multi-turn Text-to-SQL | CoSQL & SParC Aggregate | Avg Exact Match (EM)66.5 | 29 | |
| Multi-turn Text-to-SQL | SParC In-domain | Exact Match68.7 | 29 | |
| Text-to-SQL | Bird | Execution Accuracy56.9 | 20 | |
| Multi-turn Text-to-SQL | CoSQL Out-of-domain | Execution Accuracy74 | 16 | |
| Multi-turn Text-to-SQL | SParC Out-of-domain | Execution Accuracy (EX)77.4 | 16 |