Text-to-SQL on Spider-Realistic
88.6Execution Accuracy (EX)FineStep-4B
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FineStep-4BModel Size=4B2026.05 | 88.6 | — | — | — | 80.9 | |
| FineStep-8BModel Size=8B2026.05 | 86 | — | — | — | 77.9 | |
| ReEx-SQL-7BModel Size=7B2026.05 | 85.5 | — | — | — | 79.9 | |
| SENSE-13BModel Size=13B2026.05 | 84.1 | — | — | — | 76.6 | |
| SQL-R1-7B*Model Size=7B2026.05 | 83.3 | — | — | — | — | |
| SFT CodeS-15BModel Size=15B2026.05 | 83.1 | — | — | — | 75.6 | |
| SQL-o1 + Llama3-8BBackbone=Llama3-8B2026.05 | 82.7 | — | — | — | 72.8 | |
| DCG-SQLLLM Backbone=GPT-42025.05 | 81.9 | 73.6 | — | — | — | |
| Think2SQL-14BModel Scale Category=10-50B2025.04 | 81.8 | — | — | — | — | |
| ACT-SQLLLM Backbone=GPT-42025.05 | 81.3 | 55.7 | — | — | — | |
| Think2SQL-4BModel Scale Category=< 5B2025.04 | 81.2 | — | — | — | — | |
| MTIR-SQL + Qwen3-14BBackbone=Qwen3-14B2026.05 | 81.1 | — | — | — | — | |
| ASTRESLLM Backbone=GPT-42025.05 | 80.9 | 66.1 | — | — | — | |
| Think2SQL-8BModel Scale Category=5-10B2025.04 | 80.9 | — | — | — | — | |
| FastRAText+GPT-4Backbone=GPT-42026.05 | 80.9 | — | — | — | — | |
| ROUTE + Llama3-8BBackbone=Llama3-8B2026.05 | 80.9 | — | — | — | 72.6 | |
| TA-SQL + GPT-4Backbone=GPT-42026.05 | 79.5 | — | — | — | — | |
| ACE-SQLDecoding Strategy=greedy2026.06 | 79.5 | — | — | — | — | |
| DCG-SQLLLM Backbone=Deepseek-Coder-33B-Instruct2025.05 | 79.1 | 71.1 | — | — | — | |
| DCG-SQLLLM Backbone=GPT-3.5-Turbo2025.05 | 79 | 62.8 | — | — | — | |
| ASTRESLLM Backbone=GPT-3.5-Turbo2025.05 | 78 | 60.8 | — | — | — | |
| SQL-PaLM + PaLM2Backbone=PaLM22026.05 | 77.6 | — | — | — | — | |
| ASTRESLLM Backbone=Deepseek-Coder-33B-Instruct2025.05 | 77.4 | 59.3 | — | — | — | |
| MTIR-SQLModel Scale=8B, Decoding Strategy=greedy2026.06 | 77.4 | — | — | — | — | |
| OmniSQL-32BModel Size=32B2026.05 | 77.2 | — | — | — | — | |
| ACT-SQLLLM Backbone=GPT-3.5-Turbo2025.05 | 76.4 | 48.6 | — | — | — | |
| ACT-SQLLLM Backbone=Deepseek-Coder-33B-Instruct2025.05 | 76.4 | 51.8 | — | — | — | |
| OmniSQL-7BParameters=7B2026.01 | 76.2 | 61.4 | — | — | — | |
| OmniSQLModel Scale=7B, Decoding Strategy=greedy2026.06 | 76.2 | — | — | — | — | |
| DAIL-SQLLLM Backbone=GPT-42025.05 | 76 | — | — | — | — | |
| DCG-SQLLLM Backbone=Llama-3.1-8B-Instruct2025.05 | 75.8 | 67.1 | — | — | — | |
| EVOLSQL-7BParameters=7B2026.01 | 75.8 | 61.9 | — | — | — | |
| DCG-SQLLLM Backbone=Deepseek-Coder-6.7B-Instruct2025.05 | 74.8 | 68.5 | — | — | — | |
| ACT-SQLLLM Backbone=Llama-3.1-8B-Instruct2025.05 | 74.4 | 60.2 | — | — | — | |
| RASAT+PICARDPost-processing=PICARD2022.05 | 71.9 | 69.7 | — | — | — | |
| SFT (BIRD+Spider)Training Data=BIRD+Spider2026.01 | 71.9 | 55.8 | — | — | — | |
| T5-3B+PICARDPost-processing=PICARD2022.05 | 71.4 | 68.7 | — | — | — | |
| CYCLESQL + GPT-4Backbone=GPT-42026.05 | 70.6 | — | — | — | 56.9 | |
| ACT-SQLLLM Backbone=Deepseek-Coder-6.7B-Instruct2025.05 | 69.1 | 53.5 | — | — | — | |
| DCG-SQLLLM Backbone=Llama-3.2-3B-Instruct2025.05 | 68.5 | 59.6 | — | — | — | |
| DAIL-SQLLLM Backbone=GPT-3.5-Turbo2025.05 | 67.9 | — | — | — | — | |
| Base ModelBackbone=Qwen2.5-Coder-7B-Instruct2026.01 | 66.7 | 53.4 | — | — | — | |
| RASAT2022.05 | 65.8 | 65.2 | — | — | — | |
| RAT-SQL+STRUGAugmentation=STRUG2022.05 | 65.7 | 62.2 | — | — | — | |
| DIN-SQLLLM Backbone=GPT-3.5-Turbo2025.05 | 65.2 | 41.1 | — | — | — | |
| T5-3B2022.05 | 64.1 | 62 | — | — | — | |
| ACT-SQLLLM Backbone=Llama-3.2-3B-Instruct2025.05 | 62 | 47 | — | — | — | |
| DIN-SQLLLM Backbone=Llama-3.1-8B-Instruct2025.05 | 59.6 | 36.4 | — | — | — | |
| DIN-SQLLLM Backbone=Deepseek-Coder-6.7B-Instruct2025.05 | 53.5 | 29.3 | — | — | — | |
| DIN-SQLLLM Backbone=Llama-3.2-3B-Instruct2025.05 | 38.6 | 29.1 | — | — | — | |
| DataFlow-Text2SQL-50KBackbone=Meta-Llama-3.1-8B-Instruct2025.12 | — | — | 69.6 | 73.5 | — | |
| DataFlow-Text2SQL-50KBackbone=Qwen2.5-Coder-7B-Instruct2025.12 | — | — | 77.8 | 82.1 | — | |
| DataFlow-Text2SQL-90KBackbone=Meta-Llama-3.1-8B-Instruct2025.12 | — | — | 71.4 | 72.7 | — | |
| DataFlow-Text2SQL-90KBackbone=Qwen2.5-Coder-7B-Instruct2025.12 | — | — | 79.5 | 81.7 | — | |
| DeepSeek-Coder-7B-InstructAccess=Open-source2025.12 | — | — | 58.7 | 58.9 | — | |
| GPT-4-TurboAccess=Closed-source2025.12 | — | — | 67.5 | 68.3 | — | |
| GPT-4oAccess=Closed-source2025.12 | — | — | 66.5 | 66.7 | — | |
| GPT-4o-miniAccess=Closed-source2025.12 | — | — | 64.4 | 66.7 | — | |
| Granite-3.1-8B-InstructAccess=Open-source2025.12 | — | — | 46.5 | 57.1 | — | |
| Granite-8B-Code-InstructAccess=Open-source2025.12 | — | — | 48.8 | 49.4 | — | |
| Meta-Llama-3.1-8B-InstructAccess=Open-source2025.12 | — | — | 57.5 | 61 | — | |
| MTIR-SQL-4BSchema Prefill=true2026.03 | — | — | 78.7 | — | — | |
| MTIR-SQL-8BSchema Prefill=true2026.03 | — | — | 77.4 | — | — | |
| OmniSQL-7BSchema Prefill=true2026.03 | — | — | 76.2 | 78 | — | |
| OpenCoder-8B-InstructAccess=Open-source2025.12 | — | — | 49 | 49.4 | — | |
| Qwen2.5-7B-InstructAccess=Open-source2025.12 | — | — | 56.7 | 63.6 | — | |
| Qwen2.5-Coder-7B-InstructAccess=Open-source2025.12 | — | — | 66.7 | 70.5 | — | |
| Spider+BIRD+DataFlow-Text2SQL-90KBackbone=Meta-Llama-3.1-8B-Instruct2025.12 | — | — | 74.4 | 75 | — | |
| Spider+BIRD+DataFlow-Text2SQL-90KBackbone=Qwen2.5-Coder-7B-Instruct2025.12 | — | — | 82.3 | 83.7 | — | |
| SQL-R1-3BSchema Prefill=true2026.03 | — | — | — | 71.5 | — | |
| SQL-R1-7BSchema Prefill=true2026.03 | — | — | — | 83.3 | — | |
| SQL-Trail-7BSchema Prefill=true2026.03 | — | — | 79.6 | 83.9 | — | |
| SynSQL(2.5M)Backbone=Meta-Llama-3.1-8B-Instruct2025.12 | — | — | 67.9 | 69.4 | — | |
| SynSQL(2.5M)Backbone=Qwen2.5-Coder-7B-Instruct2025.12 | — | — | 76.2 | 78 | — | |
| SynSQL(50K)Backbone=Meta-Llama-3.1-8B-Instruct2025.12 | — | — | 69.3 | 71.6 | — | |
| SynSQL(50K)Backbone=Qwen2.5-Coder-7B-Instruct2025.12 | — | — | 77.2 | 80.3 | — | |
| SynSQL(90K)Backbone=Meta-Llama-3.1-8B-Instruct2025.12 | — | — | 69 | 70.9 | — | |
| SynSQL(90K)Backbone=Qwen2.5-Coder-7B-Instruct2025.12 | — | — | 77 | 79.9 | — | |
| TRUST-SQL-4BSchema Prefill=false2026.03 | — | — | 79.9 | 82.5 | — | |
| TRUST-SQL-8BSchema Prefill=false2026.03 | — | — | 82.1 | 84.1 | — |