Text-to-SQL on Spider-DK
81.9Execution Accuracy (EX)FineStep
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| FineStepModel Size=30B-A3B2026.05 | 81.9 | — | — | — | — | — | — | — | |
| FineStep-4BModel Size=4B2026.05 | 81.5 | — | — | — | — | — | — | — | |
| FineStepModel Size=4B2026.05 | 81.5 | — | — | — | — | — | — | — | |
| Arctic-Text2SQL-R1Size=32B, # agents=1, Thinking?=Yes, Training set=Bird+Spider2025.11 | 81.5 | — | — | — | — | — | — | — | |
| KE-RLLLM=Qwen2.5-Coder-7B2026.05 | 80.97 | — | — | — | — | — | — | — | |
| GRPOModel Size=4B2026.05 | 80.7 | — | — | — | — | — | — | — | |
| GIGPOModel Size=4B2026.05 | 80.4 | — | — | — | — | — | — | — | |
| SENSE-13BModel Size=13B2026.05 | 80.2 | — | — | — | — | — | — | — | |
| ReEx-SQL-7BModel Size=7B2026.05 | 80.2 | — | — | — | — | — | — | — | |
| GSPOModel Size=8B2026.05 | 80.2 | — | — | — | — | — | — | — | |
| FineStep-8BModel Size=8B2026.05 | 80 | — | — | — | — | — | — | — | |
| FineStepModel Size=8B2026.05 | 80 | — | — | — | — | — | — | — | |
| EXPO-SQLDetail=RL, Base Model=Qwen2.5-Coder-7B2026.04 | 79.9 | — | — | — | — | — | — | — | |
| DAPOModel Size=4B2026.05 | 79.6 | — | — | — | — | — | — | — | |
| DAPOModel Size=8B2026.05 | 79.4 | — | — | — | — | — | — | — | |
| GIGPOModel Size=8B2026.05 | 79.1 | — | — | — | — | — | — | — | |
| GSPOModel Size=4B2026.05 | 78.9 | — | — | — | — | — | — | — | |
| SQL-o1 + Llama3-8BBackbone=Llama3-8B2026.05 | 78.7 | — | — | — | — | — | — | — | |
| GRPOModel Size=8B2026.05 | 78.7 | — | — | — | — | — | — | — | |
| GRPOModel Size=30B-A3B2026.05 | 78.7 | — | — | — | — | — | — | — | |
| MARS-SQL2025.11 | 78.13 | — | — | — | — | — | — | — | |
| MARS-SQLSize=7B each, # agents=3, Thinking?=Yes, Training set=Bird2025.11 | 78.13 | — | — | — | — | — | — | — | |
| SQL-R1-7B*Model Size=7B2026.05 | 78.1 | — | — | — | — | — | — | — | |
| SQL-R1Detail=RL, Base Model=Qwen2.5-Coder-7B2026.04 | 78.1 | — | — | — | — | — | — | — | |
| SENSE-7BDetail=SFT, Base Model=Qwen2.5-Coder-7B2026.04 | 77.9 | — | — | — | — | — | — | — | |
| Think2SQL-14BModel Scale Category=10-50B2025.04 | 77.8 | — | — | — | — | — | — | — | |
| OmniSQL-7BDetail=SFT, Base Model=Qwen2.5-Coder-7B2026.04 | 77.8 | — | — | — | — | — | — | — | |
| OmniSQL-32BModel Size=32B2026.05 | 77.6 | — | — | — | — | — | — | — | |
| Qwen2.5-72B-InstructDecoding=Maj@82025.05 | 77.6 | — | — | — | — | — | — | — | |
| KE-RLLLM=Deepseek-Coder-7B2026.05 | 76.58 | — | — | — | — | — | — | — | |
| OmniLLM=Qwen2.5-Coder-7B, Setting=RL2026.05 | 76.52 | — | — | — | — | — | — | — | |
| Progress-SQL-14BReward=Progressive2026.06 | 76.4 | — | — | — | — | — | — | — | |
| Qwen2.5-72B-InstructDecoding=Greedy2025.05 | 76.4 | — | — | — | — | — | — | — | |
| MTIR-SQL + Qwen3-14BBackbone=Qwen3-14B2026.05 | 76.3 | — | — | — | — | — | — | — | |
| OmniSQL-7BParameters=7B2026.01 | 76.1 | — | — | — | — | — | — | 61.4 | |
| OmniSQLSize=32B, # agents=1, Thinking?=No, Training set=OmniSQL2025.11 | 76.1 | — | — | — | — | — | — | — | |
| OmniSQLModel Scale=7B, Decoding Strategy=greedy2026.06 | 76.1 | — | — | — | — | — | — | — | |
| Think2SQL-4BModel Scale Category=< 5B2025.04 | 76 | — | — | — | — | — | — | — | |
| Think2SQL-8BModel Scale Category=5-10B2025.04 | 75.9 | — | — | — | — | — | — | — | |
| Progress-SQL-7BReward=Progressive2026.06 | 75.9 | — | — | — | — | — | — | — | |
| SQL-Trail-7BReward=One-shot2026.06 | 75.4 | — | — | — | — | — | — | — | |
| KE-ICLLLM=Gemini-Pro-1.52026.05 | 75.32 | — | — | — | — | — | — | — | |
| No-TrainModel Size=30B-A3B2026.05 | 75.3 | — | — | — | — | — | — | — | |
| ShareSize=8B, # agents=1, Thinking?=No, Training set=Bird2025.11 | 75.3 | — | — | — | — | — | — | — | |
| No-TrainModel Size=8B2026.05 | 75 | — | — | — | — | — | — | — | |
| ROUTE + Llama3-8BBackbone=Llama3-8B2026.05 | 74.6 | — | — | — | — | — | — | — | |
| OmniSQL-7BReward=–2026.06 | 74.6 | — | — | — | — | — | — | — | |
| ACE-SQLDecoding Strategy=greedy2026.06 | 74.4 | — | — | — | — | — | — | — | |
| EVOLSQL-7BParameters=7B2026.01 | 74.2 | — | — | — | — | — | — | 61.9 | |
| CodeSLLM=Gemini-Pro-1.5, Setting=ICL2026.05 | 74.2 | — | — | — | — | — | — | — | |
| KE-ICLLLM=Qwen2.5-Coder-7B2026.05 | 74.02 | — | — | — | — | — | — | — | |
| SQL-GENLLM=Qwen2.5-Coder-7B, Setting=RL2026.05 | 73.9 | — | — | — | — | — | — | — | |
| DeepSeek-V3 (671B MoE)Decoding=Maj@82025.05 | 73.8 | — | — | — | — | — | — | — | |
| No-TrainModel Size=4B2026.05 | 73.6 | — | — | — | — | — | — | — | |
| Qwen2.5-Coder-7B-InstructDecoding=Maj@82025.05 | 73.6 | — | — | — | — | — | — | — | |
| GPT-4oDecoding=Maj@82025.05 | 73.5 | — | — | — | — | — | — | — | |
| Reasoning-SQL*2026.05 | 73.3 | — | — | — | — | — | — | — | |
| Reasoning-SQL-7BReward=One-shot2026.06 | 73.3 | — | — | — | — | — | — | — | |
| Reasoning-SQLDetail=RL, Base Model=Qwen2.5-Coder-7B2026.04 | 73.3 | — | — | — | — | — | — | — | |
| Reward-SQL + Qwen3-8BDecoding=Maj@82025.05 | 73.1 | — | — | — | — | — | — | — | |
| Reasoning SQLSize=14B, # agents=1, Thinking?=Yes, Training set=Bird2025.11 | 73.03 | — | — | — | — | — | — | — | |
| DAIL-SQLLLM=Gemini-Pro-1.5, Setting=ICL2026.05 | 73.02 | — | — | — | — | — | — | — | |
| TA-SQL + GPT-4Backbone=GPT-42026.05 | 72.9 | — | — | — | — | — | — | — | |
| GPT-4oThinking?=No2025.11 | 72.9 | — | — | — | — | — | — | — | |
| MTIR-SQLModel Scale=8B, Decoding Strategy=greedy2026.06 | 72.9 | — | — | — | — | — | — | — | |
| GPT-4oDecoding=Greedy2025.05 | 72.9 | — | — | — | — | — | — | — | |
| DeepSeek-V3 (671B MoE)Decoding=Greedy2025.05 | 72.9 | — | — | — | — | — | — | — | |
| OmniLLM=Deepseek-Coder-7B, Setting=RL2026.05 | 72.47 | — | — | — | — | — | — | — | |
| ASTRESLLM Backbone=GPT-42025.05 | 72.3 | 59.1 | — | — | — | — | — | — | |
| FastRAText+GPT-4Backbone=GPT-42026.05 | 72.3 | — | — | — | — | — | — | — | |
| Qwen2.5-Coder-14B-InstructReward=–2026.06 | 72.3 | — | — | — | — | — | — | — | |
| GPT-4-TurboDecoding=Greedy2025.05 | 72.3 | — | — | — | — | — | — | — | |
| GPT-4-TurboDecoding=Maj@82025.05 | 72.1 | — | — | — | — | — | — | — | |
| ACT-SQLLLM Backbone=GPT-42025.05 | 72 | 47.7 | — | — | — | — | — | — | |
| CodeSSize=7B, # agents=1, Thinking?=No, Training set=Spider2025.11 | 72 | — | — | — | — | — | — | — | |
| SFT CodeS-7BReward=–2026.06 | 72 | — | — | — | — | — | — | — | |
| SkyRL-SQL-7BReward=One-shot2026.06 | 72 | — | — | — | — | — | — | — | |
| Arctic-SQL + Qwen2.5-7BDecoding=Maj@82025.05 | 71.8 | — | — | — | — | — | — | — | |
| GPT-O3-mini2025.11 | 71.77 | — | — | — | — | — | — | — | |
| O3-miniThinking?=Yes2025.11 | 71.77 | — | — | — | — | — | — | — | |
| DCG-SQLLLM Backbone=GPT-42025.05 | 71.6 | 66.5 | — | — | — | — | — | — | |
| MAC-SQLDetail=Prompt, Base Model=GPT-42026.04 | 71.4 | — | — | — | — | — | — | — | |
| CodeSLLM=Qwen2.5-Coder-7B, Setting=ICL2026.05 | 71.29 | — | — | — | — | — | — | — | |
| SFT CodeS-15BModel Size=15B2026.05 | 70.7 | — | — | — | — | — | — | — | |
| ASTRESLLM Backbone=Deepseek-Coder-33B-Instruct2025.05 | 70.5 | 46.4 | — | — | — | — | — | — | |
| SQL-R1 + Qwen2.5-7BDecoding=Maj@82025.05 | 70.5 | — | — | — | — | — | — | — | |
| DAIL-SQLLLM=Qwen2.5-Coder-7B, Setting=ICL2026.05 | 70.43 | — | — | — | — | — | — | — | |
| SQL-GENLLM=Deepseek-Coder-7B, Setting=RL2026.05 | 70.3 | — | — | — | — | — | — | — | |
| SQL-R1-7BReward=One-shot2026.06 | 69.9 | — | — | — | — | — | — | — | |
| Meta-Llama-3.1-8B-InstructDecoding=Maj@82025.05 | 69.9 | — | — | — | — | — | — | — | |
| Arctic-SQL + Qwen2.5-7BDecoding=Greedy2025.05 | 69.7 | — | — | — | — | — | — | — | |
| ACT-SQLLLM Backbone=Deepseek-Coder-33B-Instruct2025.05 | 69.5 | 47.5 | — | — | — | — | — | — | |
| Reward-SQL + Qwen3-8BDecoding=Greedy2025.05 | 69.4 | — | — | — | — | — | — | — | |
| DCG-SQLLLM Backbone=GPT-3.5-Turbo2025.05 | 69 | 59.4 | — | — | — | — | — | — | |
| ASTRESLLM Backbone=GPT-3.5-Turbo2025.05 | 68.8 | 49.3 | — | — | — | — | — | — | |
| Qwen3-coderSize=30B, Thinking?=No2025.11 | 68.79 | — | — | — | — | — | — | — | |
| CYCLESQL + GPT-4Backbone=GPT-42026.05 | 68.5 | — | — | — | — | — | — | — | |
| KE-ICLLLM=GPT-4o2026.05 | 68.41 | — | — | — | — | — | — | — | |
| DCG-SQLLLM Backbone=Deepseek-Coder-33B-Instruct2025.05 | 68.4 | 62.8 | — | — | — | — | — | — | |
| ACT-SQLLLM Backbone=GPT-3.5-Turbo2025.05 | 68.2 | 45.4 | — | — | — | — | — | — |