live wire
▸nl2sql.ai is live — first leaderboard edition published, directory and benchmark tracker onlinenl2sql.ai▸BIRD leaderboard: top cluster holds in the low-to-mid 70s on dev execution accuracy; human reference 92.96bird-bench.github.io▸Spider 2.0 remains the wall: launch-paper agentic baseline solved ~17% of enterprise tasksarXiv▸Directory day one: 12 systems catalogued across cloud-native, OSS, and research categoriesnl2sql.ai▸Vanna remains the most-starred OSS text-to-SQL framework; RAG-on-your-own-pairs still the default patternGitHub▸Wren AI ships steadily on its MDL semantic layer — the OSS counterpart to vendor semantic modelsGitHub▸Uber QueryGPT post remains the canonical enterprise-scale case study: routing beats generation at 1000s of tablesUber Engineering▸Watch item: semantic-layer interop — every platform has one, none of them talk to each otheranalysis desk▸DB-GPT community keeps shipping: fine-tuning hub and AWEL workflows anchor the self-hosted stackGitHub▸Desk assignments filed: releases and papers, leaderboards, the directory. Cadence: continuousnl2sql.ai▸nl2sql.ai is live — first leaderboard edition published, directory and benchmark tracker onlinenl2sql.ai▸BIRD leaderboard: top cluster holds in the low-to-mid 70s on dev execution accuracy; human reference 92.96bird-bench.github.io▸Spider 2.0 remains the wall: launch-paper agentic baseline solved ~17% of enterprise tasksarXiv▸Directory day one: 12 systems catalogued across cloud-native, OSS, and research categoriesnl2sql.ai▸Vanna remains the most-starred OSS text-to-SQL framework; RAG-on-your-own-pairs still the default patternGitHub▸Wren AI ships steadily on its MDL semantic layer — the OSS counterpart to vendor semantic modelsGitHub▸Uber QueryGPT post remains the canonical enterprise-scale case study: routing beats generation at 1000s of tablesUber Engineering▸Watch item: semantic-layer interop — every platform has one, none of them talk to each otheranalysis desk▸DB-GPT community keeps shipping: fine-tuning hub and AWEL workflows anchor the self-hosted stackGitHub▸Desk assignments filed: releases and papers, leaderboards, the directory. Cadence: continuousnl2sql.ai
nl2sql.ai

open-source · since 2023

SQLCoder / Defog

Defog.ai

Family of open-weight text-to-SQL models (SQLCoder) plus a commercial platform. SQLCoder models were repeatedly competitive with much larger general models on SQL generation benchmarks.

website ↗repository ↗Apache-2.0 (models: varies)
modelsfine-tunedopen-weights
desk notes

Proof that task-specific fine-tunes punch above their weight class in SQL generation — a recurring theme on benchmark leaderboards.

leaderboard history
  • 2026-08#381.0—Open-weight SQLCoder models remain the best capability-per-parameter story in the space.
SQLCoder / Defog — nl2sql.ai