Single-hop Question Answering on LoCoMo
0.6408F1ShardMemo
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ShardMemoModel=GPT-OSS-120B2026.01 | 0.6408 | — | — | 0.597 | — | |
| Full Context2026.02 | 0.634 | 0.557 | 0.885 | — | — | |
| MEMIRBackbone=GPT-4.12026.05 | 0.622 | — | — | 0.567 | 0.868 | |
| MemboxModel=GPT-4o2026.01 | 0.6169 | 0.4936 | — | — | — | |
| LoCoMoModel=GPT-4o2026.01 | 0.6156 | — | — | 0.5419 | — | |
| LoCoMoModel=GPT-4o2026.01 | 0.6156 | 0.5419 | — | — | — | |
| LoCoMoModel=GPT-4o, Token Length=16,9102026.04 | 0.6156 | — | — | 0.5419 | — | |
| MEMGPTModel=GPT-4o2026.01 | 0.6018 | 0.5335 | — | — | — | |
| MemGPTModel=GPT-4o2026.01 | 0.6016 | — | — | 0.5335 | — | |
| MemGPTModel=GPT-4o, Token Length=16,9872026.04 | 0.6016 | — | — | 0.5335 | — | |
| MemboxModel=GPT-4o-mini2026.01 | 0.6009 | 0.4745 | — | — | — | |
| MEMORARetriever Type=Semantic Retriever2026.02 | 0.597 | 0.522 | 0.9 | — | — | |
| MEMORARetriever Type=Policy Retriever2026.02 | 0.597 | 0.521 | 0.918 | — | — | |
| MEMIRBackbone=GPT-4.1-mini2026.05 | 0.596 | — | — | 0.542 | 0.895 | |
| Nemori*2026.02 | 0.588 | 0.515 | 0.849 | — | — | |
| GAMModel=GPT-OSS-120B2026.01 | 0.5838 | — | — | 0.53 | — | |
| SimpleMemBackbone=GPT-4.1-mini2026.05 | 0.577 | — | — | 0.516 | 0.874 | |
| Nemori+AMAModel=GPT-4o-mini2026.01 | 0.5705 | — | — | 0.4646 | — | |
| NemoriModel=GPT-4o2026.01 | 0.5648 | — | — | 0.4617 | — | |
| Nemori+AMAModel=GPT-4o2026.01 | 0.5639 | — | — | 0.4603 | — | |
| LightMEMModel=GPT-4o2026.01 | 0.5588 | — | — | 0.4379 | — | |
| SimpleMemBackbone=GPT-4.12026.05 | 0.558 | — | — | 0.499 | 0.838 | |
| NEMORIBackbone=GPT-4.1-mini2026.05 | 0.557 | — | — | 0.495 | 0.87 | |
| LightMEM+AMAModel=GPT-4o2026.01 | 0.5553 | — | — | 0.4353 | — | |
| O-MemBackbone=GPT-4.12026.05 | 0.5489 | — | — | 0.4898 | — | |
| Memory-R2Training regime=Trained2026.05 | 0.5406 | — | — | 0.4873 | 0.868 | |
| NemoriModel=GPT-4o-mini2026.01 | 0.5402 | — | — | 0.4361 | — | |
| NEMORIBackbone=GPT-4.12026.05 | 0.538 | — | — | 0.454 | 0.856 | |
| HeLa-MemModel=GPT-4o-mini, Token Length=1,0102026.04 | 0.5189 | — | — | 0.4625 | — | |
| RAGModel=GPT-OSS-120B2026.01 | 0.5136 | — | — | 0.458 | — | |
| LangMemBackbone=GPT-4.12026.05 | 0.5113 | — | — | 0.4422 | — | |
| LangMem*2026.02 | 0.51 | 0.436 | 0.845 | — | — | |
| LangMemBackbone=GPT-4.1-mini2026.05 | 0.51 | — | — | 0.436 | 0.845 | |
| Memory-R2(OSS)Training regime=Trained, Answer Agent=GPT-OSS-120B2026.05 | 0.5098 | — | — | 0.4579 | 0.9058 | |
| RAG2026.02 | 0.507 | 0.448 | 0.71 | — | — | |
| MemAgentTraining regime=Trained2026.05 | 0.5055 | — | — | 0.4389 | 0.7861 | |
| Mem0Backbone=GPT-4.12026.05 | 0.504 | — | — | 0.432 | 0.759 | |
| HeLa-MemModel=Qwen2.5-14b, Token Length=9442026.04 | 0.4995 | — | — | 0.4515 | — | |
| LightMEMModel=GPT-4o-mini2026.01 | 0.4993 | — | — | 0.3917 | — | |
| LightMEM+AMAModel=GPT-4o-mini2026.01 | 0.4988 | — | — | 0.3984 | — | |
| HeLa-MemModel=GPT-4o, Token Length=1,0362026.04 | 0.4969 | — | — | 0.4408 | — | |
| HIMEMBackbone=GPT-4.1-mini2026.05 | 0.491 | — | — | 0.435 | 0.743 | |
| Mem0Backbone=GPT-4.1-mini2026.05 | 0.486 | — | — | 0.42 | 0.714 | |
| Mem0Model=GPT-4o, Local re-implementation=true, k=302026.01 | 0.4849 | 0.37 | — | — | — | |
| A-MEMModel=GPT-4o2026.01 | 0.4843 | 0.4297 | — | — | — | |
| A-MemModel=GPT-4o, Token Length=1,2162026.04 | 0.4843 | — | — | 0.4297 | — | |
| Mem0Model=GPT-4o-mini2026.01 | 0.4765 | 0.3872 | — | — | — | |
| MEMORYOSModel=GPT-OSS-120B2026.01 | 0.4717 | — | — | 0.4239 | — | |
| A-MEMModel=GPT-4o, Local re-implementation=true, k=302026.01 | 0.4704 | 0.4102 | — | — | — | |
| Mem0Model=GPT-4o-mini, Local re-implementation=true, k=302026.01 | 0.4689 | 0.3763 | — | — | — | |
| Memory-R1Training regime=Trained2026.05 | 0.4686 | — | — | 0.4092 | 0.6781 | |
| A-MemBackbone=GPT-4.12026.05 | 0.465 | — | — | 0.401 | 0.621 | |
| MemoryOSModel=Qwen2.5-14b2026.04 | 0.4633 | — | — | 0.4162 | — | |
| MemoryOSModel=GPT-4o-mini, Token Length=2,0002026.04 | 0.4586 | — | — | 0.407 | — | |
| HIMEMBackbone=GPT-4.12026.05 | 0.456 | — | — | 0.407 | 0.708 | |
| Zep*2026.02 | 0.455 | 0.4 | 0.669 | — | — | |
| ZepBackbone=GPT-4.1-mini2026.05 | 0.455 | — | — | 0.4 | 0.669 | |
| SwiftMemBackbone=GPT-4.1-mini2026.05 | 0.455 | — | — | 0.48 | 0.767 | |
| MEM0Model=GPT-OSS-120B2026.01 | 0.4535 | — | — | 0.3795 | — | |
| A-MemBackbone=GPT-4.1-mini2026.05 | 0.45 | — | — | 0.398 | 0.64 | |
| Vanilla LLMModel=GPT-OSS-120B2026.01 | 0.4472 | — | — | 0.3528 | — | |
| A-MEMModel=GPT-4o-mini2026.01 | 0.4465 | 0.3706 | — | — | — | |
| A-MemModel=GPT-4o-mini, Token Length=2,5202026.04 | 0.4465 | — | — | 0.3706 | — | |
| A-MEMModel=GPT-OSS-120B2026.01 | 0.4438 | — | — | 0.3642 | — | |
| MemoryOSBackbone=GPT-4.1-mini2026.05 | 0.442 | — | — | 0.375 | 0.689 | |
| MemoryOSModel=GPT-4o, Token Length=2,0002026.04 | 0.4385 | — | — | 0.3903 | — | |
| Mem02026.02 | 0.433 | 0.376 | 0.677 | — | — | |
| MemoryOSBackbone=GPT-4.12026.05 | 0.431 | — | — | 0.368 | 0.659 | |
| MEM0Model=Qwen2.5-14b2026.04 | 0.4258 | — | — | 0.3515 | — | |
| LIGHTMEMModel=GPT-OSS-120B2026.01 | 0.423 | — | — | 0.3777 | — | |
| MemGPTModel=GPT-4o-mini2026.01 | 0.4104 | — | — | 0.3434 | — | |
| MEMGPTModel=GPT-4o-mini2026.01 | 0.4104 | 0.3434 | — | — | — | |
| MemGPTModel=GPT-4o-mini, Token Length=16,9772026.04 | 0.4104 | — | — | 0.3434 | — | |
| MemoryOSTraining regime=Training-free2026.05 | 0.4085 | — | — | 0.3626 | 0.6314 | |
| A-MEMModel=GPT-4o-mini, Local re-implementation=true, k=302026.01 | 0.407 | 0.3263 | — | — | — | |
| A-MEM+AMAModel=GPT-4o2026.01 | 0.4037 | — | — | 0.3563 | — | |
| LoCoMoModel=GPT-4o-mini2026.01 | 0.4036 | — | — | 0.2905 | — | |
| LoCoMoModel=GPT-4o-mini2026.01 | 0.4036 | 0.2905 | — | — | — | |
| LoCoMoModel=GPT-4o-mini, Token Length=16,9102026.04 | 0.4036 | — | — | 0.2905 | — | |
| A-MEM+AMAModel=GPT-4o-mini2026.01 | 0.351 | — | — | 0.291 | — | |
| LightMemModel=Qwen2.5-14b2026.04 | 0.3492 | — | — | 0.3122 | — | |
| LightMemBackbone=GPT-4.1-mini2026.05 | 0.338 | — | — | 0.297 | 0.725 | |
| A-MemModel=Qwen2.5-14b, Token Length=1,3002026.04 | 0.3375 | — | — | 0.3004 | — | |
| Mem0Training regime=Training-free2026.05 | 0.3274 | — | — | 0.2527 | 0.652 | |
| A-MEMModel=GPT-4o2026.01 | 0.3227 | — | — | 0.2824 | — | |
| LightMemBackbone=GPT-4.12026.05 | 0.32 | — | — | 0.295 | 0.714 | |
| A-MEMTraining regime=Training-free2026.05 | 0.3058 | — | — | 0.2614 | 0.469 | |
| HeLa-MemModel=Qwen2.5-3b, Token Length=1,0722026.04 | 0.2951 | — | — | 0.2591 | — | |
| A-MEMModel=GPT-4o-mini2026.01 | 0.292 | — | — | 0.243 | — | |
| MEM1Training regime=Trained2026.05 | 0.2805 | — | — | 0.2296 | 0.5722 | |
| MemoryOSModel=Qwen2.5-3b, Token Length=2,0002026.04 | 0.2565 | — | — | 0.1878 | — | |
| A-MemModel=Qwen2.5-3b, Token Length=1,3002026.04 | 0.2363 | — | — | 0.1923 | — | |
| LoCoMoBackbone=GPT-4.1-mini2026.05 | 0.187 | — | — | 0.159 | — | |
| ReadAgentModel=GPT-4o2026.01 | 0.1246 | — | — | 0.1029 | — | |
| READAGENTModel=GPT-4o2026.01 | 0.1246 | 0.1029 | — | — | — | |
| ReadAgentModel=GPT-4o, Token Length=8052026.04 | 0.1246 | — | — | 0.1029 | — | |
| ReadAgentModel=GPT-4o-mini2026.01 | 0.0967 | — | — | 0.0766 | — | |
| READAGENTModel=GPT-4o-mini2026.01 | 0.0967 | 0.0766 | — | — | — | |
| ReadAgentModel=GPT-4o-mini, Token Length=6432026.04 | 0.0967 | — | — | 0.0766 | — | |
| ReadAgentBackbone=GPT-4.1-mini2026.05 | 0.092 | — | — | 0.079 | — |