Question Answering on RealMem
70.1LLM-Judge AccuracySEGTREEMEM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SEGTREEMEMVariant=Best, Backbone=qwen3.5-flash2026.06 | 70.1 | 36.4 | |
| SEGTREEMEMVariant=No-Propagation, Backbone=qwen3.5-flash2026.06 | 65.9 | 35.9 | |
| A-MEMVariant=Default, Backbone=qwen3.5-flash2026.06 | 63.5 | 36.1 | |
| DenseVariant=Default, Backbone=qwen3.5-flash2026.06 | 63.1 | 36.3 | |
| HippoRAGVariant=Default, Backbone=qwen3.5-flash2026.06 | 62.9 | 35.9 | |
| RAPTORVariant=Default, Backbone=qwen3.5-flash2026.06 | 61.8 | 36.2 | |
| BM25Variant=Default, Backbone=qwen3.5-flash2026.06 | 61.1 | 35.9 | |
| MemTreeVariant=Default, Backbone=qwen3.5-flash2026.06 | 59.7 | 36 | |
| Mem0Variant=Default, Backbone=qwen3.5-flash2026.06 | 48.6 | 35.1 |