EntSQL - With Evidence: leaderboard
EntSQL is an enterprise text-to-SQL benchmark of 1,066 aligned Chinese-English questions across five business domains, where the gold SQL depends on private business rules supplied as long documents. This board reports Execution accuracy (%) with concise expert-curated evidence snippets instead of the full document; higher is better.
Metric: Execution accuracy (%) with concise expert-curated evidence snippets instead of the full document; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 17.6 | #161 |
| 2 | GLM-5.1 | 17.1 | #179 |
| 3 | GPT-5.4 | 16.7 | #123 |
| 4 | Qwen 3.6 Max Preview | 16.3 | #159 |
| 5 | Claude Opus 4.6 | 15.5 | #102 |
| 6 | Gemini 3.1 Pro (Preview) | 12.8 | #94 |
| 7 | Kimi K2.6 | 10.6 | #178 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=entsql-with-evidence · How It Works · Data refreshed daily, snapshot 2026-11-13.