Text-to-SQL on Spider Real
0.841Execution AccuracyReward-SQL + Qwen3-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| Reward-SQL + Qwen3-8BDecoding=Greedy2025.05 | 0.841 | |
| Reward-SQL + Qwen3-8BDecoding=Maj@82025.05 | 0.833 | |
| Arctic-SQL + Qwen2.5-7BDecoding=Maj@82025.05 | 0.817 | |
| OmniSQL + Qwen2.5-7BDecoding=Maj@82025.05 | 0.807 | |
| SQL-R1 + Qwen2.5-7BDecoding=Maj@82025.05 | 0.805 | |
| Arctic-SQL + Qwen2.5-7BDecoding=Greedy2025.05 | 0.803 | |
| SQL-R1 + Qwen2.5-7BDecoding=Greedy2025.05 | 0.783 | |
| OmniSQL + Qwen2.5-7BDecoding=Greedy2025.05 | 0.752 | |
| Qwen2.5-Coder-7B-InstructDecoding=Maj@82025.05 | 0.705 | |
| Qwen2.5-72B-InstructDecoding=Greedy2025.05 | 0.701 | |
| Qwen2.5-72B-InstructDecoding=Maj@82025.05 | 0.685 | |
| GPT-4-TurboDecoding=Maj@82025.05 | 0.683 | |
| DeepSeek-V3 (671B MoE)Decoding=Greedy2025.05 | 0.679 | |
| GPT-4-TurboDecoding=Greedy2025.05 | 0.675 | |
| DeepSeek-V3 (671B MoE)Decoding=Maj@82025.05 | 0.669 | |
| GPT-4oDecoding=Maj@82025.05 | 0.667 | |
| Qwen2.5-Coder-7B-InstructDecoding=Greedy2025.05 | 0.667 | |
| GPT-4oDecoding=Greedy2025.05 | 0.665 | |
| Meta-Llama-3.1-8B-InstructDecoding=Maj@82025.05 | 0.61 | |
| Meta-Llama-3.1-8B-InstructDecoding=Greedy2025.05 | 0.575 |