Interactive SQL Querying on InterCode SQL
69.1Avg RewardCo-Evolving Agents
Evaluation Results
| Method | Links | |
|---|---|---|
| Co-Evolving AgentsBackbone=Qwen3-4B-Instruct-25072025.11 | 69.1 | |
| ETOBackbone=Qwen3-4B-Instruct-25072025.11 | 58.8 | |
| Co-Evolving AgentsAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 53.8 | |
| Llama-2-7B-Chat + PPOAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 52.4 | |
| Llama-2-7B-Chat + ETOAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 49.4 | |
| GPT-4Adaptation=In-context2025.11 | 38.5 | |
| GPT-3.5-TurboAdaptation=In-context2025.11 | 37.8 | |
| Llama-2-7B-Chat + RFTAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 35.6 | |
| Llama-2-7B-Chat + SFTAdaptation=Fine-tuning, Backbone=Llama-2-7B-Chat2025.11 | 30.8 | |
| SFTBackbone=Qwen3-4B-Instruct-25072025.11 | 12.7 |