UniQL - All-Dialect Consistency: leaderboard
Cross-dialect consistency on UniQL (2026), a 16-dialect text-to-SQL benchmark built from BIRD: the percent of the 1,534 questions that a model answers correctly, by execution match, in all 16 SQL dialects. Higher is better.
Metric: Share of the 1,534 UniQL questions (%) answered correctly (execution match) in all 16 SQL dialects; inference-only, one run; higher is better. Source: arxiv.org. Saturation forecast: Around 2030. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 20.14 | #250 |
| 2 | GPT-5 Mini | 17.99 | #348 |
| 3 | Qwen 3 32B | 14.41 | #960 |
| 4 | GPT-5.1 Codex | 14.34 | #288 |
| 5 | Llama 3 70B Instruct | 14.08 | #1676 |
| 6 | DeepSeek V4 Flash | 13.89 | #258 |
| 7 | Qwen 3 8B | 12.06 | #1830 |
| 8 | Gemini 2.5 Pro | 9.19 | #266 |
| 9 | GPT-3.5 Turbo | 9.13 | #2753 |
| 10 | Qwen 3 4B | 8.8 | #2281 |
| 11 | Qwen 3 1.7B | 5.8 | #5520 |
| 12 | Llama 3 8B Instruct | 2.54 | #4612 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=uniql-all-dialect-consistency · How It Works · Data refreshed daily, snapshot 2026-11-13.