Long-context Question Answering on LOCOMO single
68.62F1 ScoreMRAgent
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MRAgentModel (Backbone)=Claude2026.06 | 68.62 | 91.1 | |
| MRAgentModel (Backbone)=Gemini2026.06 | 64.08 | 90.48 | |
| A-MemModel (Backbone)=Claude2026.06 | 55.19 | 74.71 | |
| MemoryOSModel (Backbone)=Gemini2026.06 | 54.82 | 71.9 | |
| Mem0Model (Backbone)=Claude2026.06 | 54.43 | 74.07 | |
| Mem0Model (Backbone)=Gemini2026.06 | 54.37 | 73.72 | |
| LangMemModel (Backbone)=Claude2026.06 | 54.36 | 83.12 | |
| RAGModel (Backbone)=Gemini2026.06 | 53.69 | 69.2 | |
| RAGModel (Backbone)=Claude2026.06 | 53.66 | 69.2 | |
| LangMemModel (Backbone)=Gemini2026.06 | 48.2 | 69.68 | |
| A-MemModel (Backbone)=Gemini2026.06 | 46.39 | 61.83 | |
| MemoryOSModel (Backbone)=Claude2026.06 | 45.46 | 66.49 | |
| Zep2026.01 | 35.25 | 52.48 | |
| CAST2026.01 | 33.94 | 54.63 | |
| LangMem2026.01 | 31.45 | 30.75 | |
| HippoRAG2026.01 | 30.4 | 54.26 | |
| Mem02026.01 | 25.08 | 41.49 | |
| RAG2026.01 | 20.95 | 36.52 |