---
title: "LuthorSQL: A One-Pass Text-to-SQL Model... | SI Paper Desk"
url: https://thesimind.si/papers/2611.17780/
description: "A small local 7B model can answer nearly half of database questions in one fast pass and hand the hard ones to a big SI agent, cutting API cost by 33.3% at..."
---

Number 4 of 12 · November 16, 2026

# LuthorSQL: A One-Pass Text-to-SQL Model as System 1 for a Slower LLM Agent

A small local 7B model can answer nearly half of database questions in one fast pass and hand the hard ones to a big SI agent, cutting API cost by 33.3% at the same accuracy.

Ibrahim Memon, Armin Karamzade, M. Hanif Jhaveri and 1 more · Submitted November 13, 2026 · arXiv 2611.17780v1 · cs.CL

[Read the paper on arXiv](https://arxiv.org/abs/2611.17780) · [The PDF](https://arxiv.org/pdf/2611.17780) · [The day's 12](https://thesimind.si/papers/day/2026-11-16/)

## What they found

The authors built LuthorSQL, a 7B model that turns plain questions into database queries (SQL) in a single forward pass instead of writing text word by word. It scores slots in a fixed grammar of query plans, runs up to 32 candidate plans, and picks one with a trained validator. Alone, it reaches 68.03% execution accuracy on BIRD's hidden test set at 0.84 s per question on a single GPU. It keeps its accuracy in 9 GB with 8-bit quantized weights, and it transfers zero-shot to Snowflake, BigQuery, MySQL, PostgreSQL, and DuckDB within 3.1 points of its SQLite accuracy. Paired with a Claude Opus 5.5 agent in a pre-registered confidence cascade over 227 questions on databases published after Opus's training cutoff, it answers nearly half of the questions locally and cuts API cost by 33.3% at the same accuracy. The authors also report that Opus reproduces 48-92% of public benchmark gold queries verbatim without seeing the schema, yet its tool-assisted accuracy holds on renamed schemas and reaches 95.6-97.4% on post-cutoff databases.

## Why it matters

Organizations that want to query their data with SI agents face slow, costly calls to hosted models, and their sensitive queries leave the building. This work shows a way to keep many of those queries on hardware they already own. The contamination results also matter for how we read benchmark scores, since a model may have seen the answers. The abstract reports the cost cut on only 227 questions, so how well it holds elsewhere is not stated.

## How it changes the race

Fast small models that decide when to call a big agent could lower the cost and delay of deploying frontier SI on private data. This is a practical engineering gain in cost and privacy, not a jump in raw capability.

## Safety and control

Keeping many queries local reduces exposure of sensitive organizational data to external APIs. The contamination probing is also a useful check that benchmark scores reflect real skill and not memorized answers.

### The abstract, by the authors

Frontier LLM agents that query databases with tools are among the most accurate text-to-SQL systems, but per-query calls to a hosted model are slow, costly, and expose sensitive organizational queries to external APIs. For example, a Claude Opus 5.5 agent serving 10,000 queries a day costs over $70,000 a year at list price. To what extent can local hardware already owned by an organization offload this burden without compromising the agent's accuracy? We present LuthorSQL, a non-autoregressive 7B text-to-SQL model that serves as a rapid System 1 gate for a slower LLM agent (System 2), escalating queries only when uncertain. LuthorSQL makes a single forward pass to predict slot scores over a bounded grammar for query plans, compiles and runs up to 32 candidate plans, and chooses an answer using a trained validator. On its own, LuthorSQL reaches 68.03% execution accuracy on BIRD's hidden test set at 0.84 s per question on a single GPU, maintains its accuracy in 9 GB with 8-bit quantized weights, and transfers zero-shot to Snowflake, BigQuery, MySQL, PostgreSQL, and DuckDB within 3.1 points of its native SQLite accuracy. Combined with Claude Opus 5.5 in a pre-registered confidence cascade over 227 questions on databases published after Opus's training cutoff, LuthorSQL answers nearly half of all questions locally and cuts API cost by 33.3% at the same accuracy. Finally, contamination probing shows that Opus reproduces 48-92% of public benchmark gold queries verbatim without seeing the schema; however, its tool-assisted accuracy holds on renamed schemas and reaches 95.6-97.4% on post-cutoff databases. Code and data: https://github.com/IAMIbrahimmemon/luthorsql

Quoted from arXiv:2611.17780. The research and its words belong to its authors.

Desk score **82** out of 100

Why it was picked

- Agents (in the title)
- Interpretability
- Compute
- Evals
- Language models (in the title)
- Cross-listed in 4 fields
- Studies frontier models (Claude)

Topics

- [Efficiency](https://thesimind.si/papers/?tag=efficiency)
- [Agents](https://thesimind.si/papers/?tag=agents)
- [Language models](https://thesimind.si/papers/?tag=language-models)
- [Privacy](https://thesimind.si/papers/?tag=privacy)
- [Benchmarks](https://thesimind.si/papers/?tag=benchmarks)
- [Evals](https://thesimind.si/papers/?tag=evals)

**arXiv ID:** 2611.17780v1
**Submitted:** November 13, 2026
**Fields:** cs.CL (Computation and Language), cs.AI (Artificial Intelligence), cs.DB (Databases), cs.LG (Machine Learning)
**Level:** Intermediate
**Comments:** 8-page main text plus appendix, 23 pages

Fingerprint, SHA-256

`720724afecdc8e416ad3cded1e8ce4be0dae38f0173ee161c9f6c0d65631112f`

On Solana: [Proof of Mind seal, 3mXUMW...GSim](https://solscan.io/tx/3mXUMW5MkdFiyPBNPYd5DfwHhma5EdjhUFEw9jZaDuuPqgszJtoDYttSrD9kyWUFdMCzn7RnFNd4nQhwAUgrGSim)

### The exact text fingerprinted

{"arxiv":"2611.17780","version":1,"title":"LuthorSQL: A One-Pass Text-to-SQL Model as System 1 for a Slower LLM Agent","authors":["Ibrahim Memon","Armin Karamzade","M. Hanif Jhaveri","Roy Fox"],"day":"2026-11-16","rank":4,"score":82,"status":"written","summary":{"takeaway":"A small local 7B model can answer nearly half of database questions in one fast pass and hand the hard ones to a big SI agent, cutting API cost by 33.3% at the same accuracy.","found":"The authors built LuthorSQL, a 7B model that turns plain questions into database queries (SQL) in a single forward pass instead of writing text word by word. It scores slots in a fixed grammar of query plans, runs up to 32 candidate plans, and picks one with a trained validator. Alone, it reaches 68.03% execution accuracy on BIRD's hidden test set at 0.84 s per question on a single GPU. It keeps its accuracy in 9 GB with 8-bit quantized weights, and it transfers zero-shot to Snowflake, BigQuery, MySQL, PostgreSQL, and DuckDB within 3.1 points of its SQLite accuracy. Paired with a Claude Opus 5.5 agent in a pre-registered confidence cascade over 227 questions on databases published after Opus's training cutoff, it answers nearly half of the questions locally and cuts API cost by 33.3% at the same accuracy. The authors also report that Opus reproduces 48-92% of public benchmark gold queries verbatim without seeing the schema, yet its tool-assisted accuracy holds on renamed schemas and reaches 95.6-97.4% on post-cutoff databases.","matters":"Organizations that want to query their data with SI agents face slow, costly calls to hosted models, and their sensitive queries leave the building. This work shows a way to keep many of those queries on hardware they already own. The contamination results also matter for how we read benchmark scores, since a model may have seen the answers. The abstract reports the cost cut on only 227 questions, so how well it holds elsewhere is not stated.","race":"Fast small models that decide when to call a big agent could lower the cost and delay of deploying frontier SI on private data. This is a practical engineering gain in cost and privacy, not a jump in raw capability.","safety":"Keeping many queries local reduces exposure of sensitive organizational data to external APIs. The contamination probing is also a useful check that benchmark scores reflect real skill and not memorized answers.","level":"intermediate","tags":["efficiency","agents","language-models","privacy","benchmarks","evals"]}}

[Check it on Verify a thought](https://thesimind.si/proof-of-mind/)

## The authors

Ibrahim Memon, Armin Karamzade, M. Hanif Jhaveri, Roy Fox

Credit for this research belongs to its authors. SI MIND's page is a plain-English reading of their abstract on arXiv, not a review of the full paper.

Related on the desk

## More papers on these topics

[Nov 12, 2026 · Number 2AgentScript-Eval: Benchmarking LLMs and Agents for Code Generation in an Enterprise DSLOn a new benchmark for a private Salesforce agent language, SI models mostly fail zero-shot, and the setup around the model, like a compile-and-edit loop and clear guidance, mattered more than model size.BenchmarksCodeAgents](https://thesimind.si/papers/2611.13183/) · [Nov 11, 2026 · Number 10AMSchem: Automated Schematic Visualization of AMS Circuits via Verifiable Agentic GenerationAMSchem uses an SI agent with strict automatic checks to turn analog circuit netlists into readable schematics, reaching 85.0% validity versus 32.5% for direct generation on a 40-circuit subset.AgentsCodeEvals](https://thesimind.si/papers/2611.12333/) · [Nov 12, 2026 · Number 9SLIC: Reinforcement Fine-Tuning Small LMs for Multi-Turn Analog Circuit OptimizationA small 4B-scale SI model, trained with reinforcement fine-tuning and simulator feedback, beats a much larger hosted model at tuning analog circuits and can run locally on proprietary chip data.Reinforcement learningLanguage modelsAgents](https://thesimind.si/papers/2611.13452/)

[Number 3](https://thesimind.si/papers/2611.17610/) · [All 12 for Nov 16, 2026](https://thesimind.si/papers/day/2026-11-16/) · [Number 5](https://thesimind.si/papers/2611.17564/)
