List your company

Text-to-SQL on Spider Realistic (EX, TS)

88.6Execution Accuracy (EX)FineStep
44.455.87567.3578.825Apr 29, 2026May 5, 2026May 11, 2026May 17, 2026May 23, 2026May 29, 2026Jun 5, 2026

Evaluation results

MethodLinks
FineStepModel Size=4B2026.0588.680.9
GIGPOModel Size=4B2026.0587.680.5
DAPOModel Size=4B2026.0587.278.5
GSPOModel Size=4B2026.058678.1
FineStepModel Size=8B2026.058677.9
GIGPOModel Size=8B2026.0585.677
DAPOModel Size=8B2026.0585.476
GSPOModel Size=8B2026.0584.877
FineStepModel Size=30B-A3B2026.0584.176
Progress-SQL-7BReward=Progressive2026.0684.176.4
Progress-SQL-14BReward=Progressive2026.0684.176.6
GRPOModel Size=4B2026.0583.775.2
GRPOModel Size=8B2026.0583.775.8
SQL-Trail-7BReward=One-shot2026.0683.774.4
EXPO-SQLDetail=RL, Base Model=Qwen2.5-Coder-7B2026.0483.4—
GRPOModel Size=30B-A3B2026.0583.373.6
SQL-R1Detail=RL, Base Model=Qwen2.5-Coder-7B2026.0483.3—
SFT CodeS-7BReward=–2026.0682.977.2
SENSE-7BDetail=SFT, Base Model=Qwen2.5-Coder-7B2026.0482.7—
No-TrainModel Size=4B2026.0582.566.3
No-TrainModel Size=30B-A3B2026.0582.564.6
Qwen2.5-Coder-14B-InstructReward=–2026.0681.574.6
SkyRL-SQL-7BReward=One-shot2026.0681.1—
No-TrainModel Size=8B2026.0580.965.4
OmniSQL-7BReward=–2026.0680.575
KE-ICLLLM=Gemini-Pro-1.52026.0580.11—
SQL-R1-7BReward=One-shot2026.0680.161.6
MAC-SQLDetail=Prompt, Base Model=GPT-42026.0479.9—
CodeSLLM=Gemini-Pro-1.5, Setting=ICL2026.0579.72—
DAIL-SQLLLM=Gemini-Pro-1.5, Setting=ICL2026.0578.3—
KE-ICLLLM=GPT-4o2026.0578.14—
OmniSQL-7BDetail=SFT, Base Model=Qwen2.5-Coder-7B2026.0478—
KE-RLLLM=Qwen2.5-Coder-7B2026.0577.58—
CodeSLLM=GPT-4o, Setting=ICL2026.0576.77—
ACT-SQLDetail=Prompt, Base Model=GPT-42026.0475.8—
DAIL-SQLLLM=GPT-4o, Setting=ICL2026.0575.51—
Qwen2.5-Coder-7BDetail=-, Base Model=Qwen2.5-Coder-7B2026.0475.4—
Qwen2.5-Coder-7B-InstructReward=–2026.0674.466.9
OmniLLM=Qwen2.5-Coder-7B, Setting=RL2026.0572.43—
KE-ICLLLM=Qwen2.5-Coder-7B2026.0571.45—
SQL-GENLLM=Qwen2.5-Coder-7B, Setting=RL2026.0570.35—
KE-RLLLM=Deepseek-Coder-7B2026.0568.73—
KE-ICLLLM=Deepseek-Coder-7B2026.0566.2—
OmniLLM=Deepseek-Coder-7B, Setting=RL2026.0564.85—
CodeSLLM=Qwen2.5-Coder-7B, Setting=ICL2026.0564.8—
DAIL-SQLLLM=Qwen2.5-Coder-7B, Setting=ICL2026.0563.71—
SQL-GENLLM=Deepseek-Coder-7B, Setting=RL2026.0562.72—
KE-RLLLM=Granite-3.1-8B2026.0560.93—
CodeSLLM=Deepseek-Coder-7B, Setting=ICL2026.0557.13—
SQL-GENLLM=Granite-3.1-8B, Setting=RL2026.0557.01—
DAIL-SQLLLM=Deepseek-Coder-7B, Setting=ICL2026.0556.02—
OmniLLM=Granite-3.1-8B, Setting=RL2026.0555.1—
KE-ICLLLM=Granite-3.1-8B2026.0550.13—
CodeSLLM=Granite-3.1-8B, Setting=ICL2026.0547.43—
DAIL-SQLLLM=Granite-3.1-8B, Setting=ICL2026.0546.1—