Long-context Memory Retrieval and Reasoning on LoCoMo 32K
46.39F1 ScoreMemSifter
Evaluation Results
| Method | Links | |
|---|---|---|
| MemSifterBackbone LLM=Qwen3-30B-A3B-Ins2026.03 | 46.39 | |
| HippoRAGBackbone LLM=Qwen3-30B-A3B-Ins2026.03 | 41.94 | |
| MemSifterBackbone LLM=DeepSeek V3.22026.03 | 41.79 | |
| ReasonRankBackbone LLM=Qwen3-30B-A3B-Ins2026.03 | 41.04 | |
| NemoriBackbone LLM=Qwen3-30B-A3B-Ins2026.03 | 40.96 | |
| Qwen3-30BBackbone LLM=Qwen3-30B-A3B-Ins2026.03 | 39.81 | |
| RearankBackbone LLM=Qwen3-30B-A3B-Ins2026.03 | 39.19 | |
| BGE-M3Backbone LLM=Qwen3-30B-A3B-Ins2026.03 | 38.73 | |
| GemmaEmbBackbone LLM=Qwen3-30B-A3B-Ins2026.03 | 36.25 | |
| Mem0Backbone LLM=Qwen3-30B-A3B-Ins2026.03 | 36.2 | |
| A-MEMBackbone LLM=Qwen3-30B-A3B-Ins2026.03 | 35.87 | |
| RearankBackbone LLM=DeepSeek V3.22026.03 | 33.04 | |
| ReasonRankBackbone LLM=DeepSeek V3.22026.03 | 31.94 | |
| A-MEMBackbone LLM=DeepSeek V3.22026.03 | 31.82 | |
| Mem0Backbone LLM=DeepSeek V3.22026.03 | 31.62 | |
| DS-V3.2Backbone LLM=DeepSeek V3.22026.03 | 30.97 | |
| GemmaEmbBackbone LLM=DeepSeek V3.22026.03 | 29.9 | |
| BGE-M3Backbone LLM=DeepSeek V3.22026.03 | 29.49 | |
| HippoRAGBackbone LLM=DeepSeek V3.22026.03 | 28.11 | |
| NemoriBackbone LLM=DeepSeek V3.22026.03 | 27.43 |