---
id: 20261117-T0-08
title: "LuthorSQL：一次生成SQL的轻量模型，为慢速LLM代理做前置筛选"
title_en: "LuthorSQL: One-Pass Text-to-SQL Model Acts as Fast Filter for LLM Agents"
url: https://ai.daily.yangsir.net/daily/20261117-T0-08
issue_date: 2026-11-17
publish_date: 2026-11-16T05:00:00.000Z
category: research
source_name: "arXiv cs.CL (NLP)"
source_url: https://arxiv.org/abs/2611.17780
---

# LuthorSQL：一次生成SQL的轻量模型，为慢速LLM代理做前置筛选

研究者提出LuthorSQL，一个一次生成SQL的文本到SQL模型，定位为慢速LLM代理的"系统1"。前沿LLM代理用工具查询数据库时准确率高，但每次调用托管模型都慢、贵，且会把敏感查询暴露给外部API。LuthorSQL的做法是先用轻量模型快速出一版SQL，减少对大型模型的调用次数。对需要在内部数据库上跑text-to-SQL的团队，这提供了一种降低延迟和成本的思路。

## English Version

**LuthorSQL: One-Pass Text-to-SQL Model Acts as Fast Filter for LLM Agents**

Researchers introduced LuthorSQL, a one-pass text-to-SQL model designed as a fast "System 1" for slower LLM agents. Frontier LLM agents that query databases via tools are accurate but slow and costly per call, and they expose sensitive queries to external APIs. LuthorSQL generates a SQL draft quickly with a lightweight model, reducing calls to the larger model. For teams running text-to-SQL on internal databases, it offers a path to lower latency and cost.

---

**来源**：[arXiv cs.CL (NLP)](https://arxiv.org/abs/2611.17780)

**详情页**：https://ai.daily.yangsir.net/daily/20261117-T0-08

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*