Graph Reasoning on GRBENCH E-Commerce
0.585QwenScoreGraphScout
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GraphScoutLLM=8B w/ train2026.03 | 0.585 | 56.9 | |
| GraphScoutLLM=4B w/ train2026.03 | 0.575 | 56.2 | |
| CypherLLM=DeepSeek-Chat2026.03 | 0.545 | 11.4 | |
| CypherLLM=Qwen-Max2026.03 | 0.52 | 11.7 | |
| GraphCounselorLLM=DeepSeek-Chat2026.03 | 0.51 | 48.9 | |
| CypherLLM=GLM-4.62026.03 | 0.49 | 11.4 | |
| GraphCounselorLLM=GPT-4o2026.03 | 0.49 | 38.1 | |
| GraphCounselorLLM=GLM-4.62026.03 | 0.49 | 45.9 | |
| CypherLLM=GPT-4o2026.03 | 0.485 | 10.1 | |
| GraphCoTLLM=DeepSeek-Chat2026.03 | 0.47 | 41.8 | |
| GraphCoTLLM=Qwen-Max2026.03 | 0.465 | 41.8 | |
| GraphCounselorLLM=Qwen-Max2026.03 | 0.459 | 44.7 | |
| PolyGLLM=GLM-4.62026.03 | 0.445 | 9 | |
| PolyGLLM=DeepSeek-Chat2026.03 | 0.445 | 8.7 | |
| PolyGLLM=GPT-4o2026.03 | 0.44 | 9.2 | |
| PolyGLLM=Qwen-Max2026.03 | 0.43 | 9.8 | |
| GraphCoTLLM=GPT-4o2026.03 | 0.41 | 36 | |
| GraphCoTLLM=GLM-4.62026.03 | 0.38 | 38.8 | |
| GraphScoutLLM=8B w/o train2026.03 | 0.365 | 35.8 | |
| GraphRAGLLM=GLM-4.62026.03 | 0.325 | 35 | |
| GraphRAGLLM=GPT-4o2026.03 | 0.315 | 30.8 | |
| TextRAGLLM=GLM-4.62026.03 | 0.25 | 24.1 | |
| GraphScoutLLM=4B w/o train2026.03 | 0.25 | 25.5 | |
| TextRAGLLM=GPT-4o2026.03 | 0.2 | 18.1 | |
| GraphRAGLLM=Qwen-Max2026.03 | 0.18 | 24.3 | |
| GraphRAGLLM=DeepSeek-Chat2026.03 | 0.165 | 16.7 | |
| TextRAGLLM=DeepSeek-Chat2026.03 | 0.115 | 13.8 | |
| BaseLLMLLM=GPT-4o2026.03 | 0.11 | 8 | |
| BaseLLMLLM=DeepSeek-Chat2026.03 | 0.095 | 11.5 | |
| TextRAGLLM=Qwen-Max2026.03 | 0.09 | 14.8 | |
| BaseLLMLLM=GLM-4.62026.03 | 0.06 | 11.1 | |
| BaseLLMLLM=Qwen-Max2026.03 | 0.06 | 10.5 |