Question Answering on LoCoMo Temporal
59.1F1 ScoreDeltaMem-4o-mini
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DeltaMem-4o-miniTraining Protocol=Training-free, Memory Manager Architecture=Agentic, Backbone=GPT-4o-mini2026.04 | 59.1 | 49.7 | 73.21 | |
| DeltaMem-8B-RLTraining Protocol=RL-trained, Memory Manager Architecture=Agentic, Backbone=Qwen3-8B2026.04 | 58.79 | 47.61 | 73.83 | |
| DeltaMem-4B-RLTraining Protocol=RL-trained, Memory Manager Architecture=Agentic, Backbone=Qwen3-4B2026.04 | 57.75 | 46.29 | 72.59 | |
| DeltaMem-8BTraining Protocol=Training-free, Memory Manager Architecture=Agentic, Backbone=Qwen3-8B2026.04 | 57.66 | 44.54 | 69.78 | |
| DeltaMem-4BTraining Protocol=Training-free, Memory Manager Architecture=Agentic, Backbone=Qwen3-4B2026.04 | 56.73 | 45.26 | 70.09 | |
| TA-MemToken Usage=37552026.03 | 55.95 | 51.47 | — | |
| Mem0Training Protocol=Training-free, Memory Manager Architecture=Multi-Agent2026.04 | 54.21 | 44.53 | 59.5 | |
| LightMemTraining Protocol=Training-free, Memory Manager Architecture=Multi-Agent2026.04 | 53.94 | 39.93 | 69.16 | |
| Mem0Token Usage=17642026.03 | 48.93 | 40.51 | — | |
| A-MemToken Usage=25202026.03 | 45.85 | 36.67 | — | |
| ZepTraining Protocol=Training-free, Memory Manager Architecture=Multi-Agent2026.04 | 42 | 34.53 | 49.31 | |
| MemoryOSToken Usage=38742026.03 | 41.15 | 30.76 | — | |
| Memory-R1-7BTraining Protocol=RL-trained, Memory Manager Architecture=Multi-Agent, Backbone=7B2026.04 | 40.96 | 23.55 | 69.16 | |
| A-MemTraining Protocol=Training-free, Memory Manager Architecture=Multi-Agent2026.04 | 35.4 | 31.08 | 49.91 | |
| LangMemTraining Protocol=Training-free, Memory Manager Architecture=Multi-Agent2026.04 | 30.75 | 25.84 | 23.43 | |
| MemGPTToken Usage=169772026.03 | 25.52 | 19.44 | — | |
| LoCoMoToken Usage=169102026.03 | 18.41 | 14.77 | — | |
| ReadAgentToken Usage=6432026.03 | 12.6 | 8.87 | — | |
| MemoryBankToken Usage=4322026.03 | 9.68 | 6.99 | — |