OpenLM Text2SQL - Spider EM: leaderboard
Exact-match accuracy on Spider's cross-domain text-to-SQL benchmark, comparing predicted SQL structure clause-by-clause against gold queries across 200+ databases.
Metric: Exact Match (%). Source: openlm.ai. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MiniSeek | 81.5 | |
| 2 | Graphix-3B + PICARD | 74 | |
| 3 | CatSQL + GraPPa | 73.9 | |
| 4 | SHiP + PICARD | 73.1 | |
| 5 | G³R + LGESQL + ELECTRA | 72.9 | |
| 6 | RESDSQL+T5-1.1-lm100k-xl | 72.4 | |
| 7 | T5-SR | 72.4 | |
| 8 | N-best List Rerankers + PICARD | 72.2 | |
| 9 | S²SQL + ELECTRA | 72.1 | |
| 10 | RESDSQL-3B + NatSQL | 72 |
Interactive version: theaggregate.ai/benchmark?slug=openlm-text2sql-spider-em · How It Works · Data refreshed daily, snapshot 2026-10-20.