DBLifeBench - SQL Debugging (Single-Round): leaderboard
SQL debugging in DBLifeBench: given a failing SQLite statement and its error message, the model must produce a corrected statement in one attempt, judged by execution; percent solved, higher is better.
Metric: Debugging success (%; the corrected SQL executes correctly after one attempt, 523 erroneous statements with their SQLite error messages; temperature 0.3, top-p 0.2, SQLite). Source: arxiv.org. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 68.39 | #741 |
| 2 | GPT-4o Mini | 60.44 | #1623 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=dblifebench-sql-debugging-single-round · How It Works · Data refreshed daily, snapshot 2026-11-13.