Answer Generation on CORAL
28.9F1Ours
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| OursLLM=Qwen-2.5-7b2026.01 | 28.9 | 46.8 | |
| ClaudeLLM=Sonnet-3.52026.01 | 27.4 | — | |
| ChatGPTLLM=GPT-3.52026.01 | 26.8 | — | |
| EvoRAGLLM=Qwen-2.5-7b2026.01 | 25.1 | — | |
| UniConvLLM=Mistral-2-7b2026.01 | 24.3 | — | |
| AgenticLMLLM=Qwen-3-8b2026.01 | 24.1 | 44.7 | |
| OursLLM=Qwen-2.5-3b2026.01 | 22.4 | 43.2 | |
| AgenticLMLLM=Qwen-3-4b2026.01 | 22.1 | 42.6 | |
| ChatQALLM=LLaMA-3-8b2026.01 | 20.3 | — | |
| SFTLLM=Qwen-2.5-7b2026.01 | 18.8 | 43 | |
| SFTLLM=Qwen-2.5-3b2026.01 | 15.2 | 42.3 | |
| Search-R1LLM=Qwen-2.5-3b2026.01 | 3.9 | 41.2 | |
| Search-R1LLM=Qwen-2.5-7b2026.01 | 3.8 | 43 |