SCALE SQL Dialect Conversion — leaderboard
Metric: Ability Score (%). Source: sql-llm-leaderboard.com. 35 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.7 Max | 77.3 |
| 2 | Gemini 2.5 Pro | 77.1 |
| 3 | Gemini 3 Pro | 77.1 |
| 4 | Claude Sonnet 4.6 | 76.7 |
| 5 | Gemini 3.5 Flash | 75.8 |
| 6 | Gemini 2.5 Flash | 75.3 |
| 7 | GPT-5 | 74.9 |
| 8 | O4 Mini (High) | 74.7 |
| 9 | DeepSeek R1 | 74.5 |
| 10 | Claude Opus 4.7 | 74.4 |
| 11 | Claude Opus 4.8 | 74 |
| 12 | Claude Haiku 4.5 | 73.3 |
| 13 | Claude Sonnet 4.5 | 72.2 |
| 14 | GPT-5.5 | 71.1 |
| 15 | Qwen 3.6 35B A3B | 70.8 |
Interactive version: aibenchmarks.dev/benchmark?slug=scale-sql-dialect-conversion · How the rankings work · Data refreshed daily, snapshot 2026-07-19.