Long-context Question Answering on LOCOMO Temporal
69.82F1 ScoreMRAgent
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MRAgentModel (Backbone)=Claude2026.06 | 69.82 | 85.34 | |
| MRAgentModel (Backbone)=Gemini2026.06 | 67.66 | 80.37 | |
| Mem0Model (Backbone)=Gemini2026.06 | 58.19 | 61.68 | |
| LangMemModel (Backbone)=Claude2026.06 | 56.64 | 80.68 | |
| Mem0Model (Backbone)=Claude2026.06 | 49.5 | 53.58 | |
| A-MemModel (Backbone)=Claude2026.06 | 47.73 | 55.48 | |
| LangMemModel (Backbone)=Gemini2026.06 | 44.7 | 53.58 | |
| RAGModel (Backbone)=Gemini2026.06 | 43.52 | 49.22 | |
| RAGModel (Backbone)=Claude2026.06 | 43.39 | 48.29 | |
| A-MemModel (Backbone)=Gemini2026.06 | 40.22 | 49.53 | |
| MemoryOSModel (Backbone)=Claude2026.06 | 39.14 | 51.09 | |
| MemoryOSModel (Backbone)=Gemini2026.06 | 35.91 | 47.04 |