UniQL - Oracle: leaderboard

Metric: Execution accuracy (%) on the Oracle dialect of UniQL (1,534 BIRD development-set questions with human-verified SQL in 16 aligned dialects): given the target dialect, schema and question, the model writes one query whose execution result must match the gold result; inference-only, one run; higher is better. Source: arxiv.org. Saturation forecast: Around 2031. 13 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.563.75
2GPT-5 Mini60.56
3DeepSeek V4 Flash57.5
4Qwen 3 32B53.06
5Llama 3 70B Instruct51.56
6Qwen 3 8B51.11
7GPT-5.1 Codex49.67
8GPT-3.5 Turbo49.35
9Qwen 3 4B44.92
10Qwen 3 1.7B43.48
11Gemini 2.5 Pro34.29
12Llama 3 8B Instruct31.23

Interactive version: theaggregate.ai/benchmark?slug=uniql-oracle · How It Works · Data refreshed daily, snapshot 2026-09-29.