# Long-term dialogue memory on LoCoMo (test) Source: [Long-term dialogue memory on LoCoMo (test)]() Collection index: [llms.txt]() Data updated: 2026-04-24T08:44:01.887Z Evaluation of model performance on the test split of the LoCoMo dataset specifically measuring long-term dialogue memory capabilities using accuracy metrics. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Long-term dialogue memory | | Dataset (including any stated split) | LoCoMo (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 30 of 30 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy](): Higher is better; unit: PERCENT. - [Accuracy (Single-Hop)](): Higher is better; unit: PERCENT. - [Accuracy (Multi-Hop)](): Higher is better; unit: PERCENT. - [Accuracy (Open-Domain)](): Higher is better; unit: PERCENT. - [Accuracy (Temporal)](): Higher is better; unit: PERCENT. - [Wrong Answers Count](): Lower is better; unit: SCALAR. - [Accuracy (Actual)](): Higher is better; unit: PERCENT. - [Utility Score](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | MemBuilder | 84.23 | Not recorded | 69647fe753fdd5c361744743 | | MemBuilder | 82.61 | Not recorded | 69647fe753fdd5c361744743 | | MemBuilder | 81.74 | Not recorded | 69647fe753fdd5c361744743 | | MemBuilder | 79.91 | Not recorded | 69647fe753fdd5c361744743 | | MemBuilder | 79.31 | Not recorded | 69647fe753fdd5c361744743 | | MIRIX | 77.48 | Not recorded | 69647fe753fdd5c361744743 | | MemBuilder | 77.47 | Not recorded | 69647fe753fdd5c361744743 | | MIRIX | 77.37 | Not recorded | 6996990909c33900fcd71efc | | StructMem | 76.82 | Not recorded | 69eaf94055a3c32af89f8b75 | | MMA | 75.94 | Not recorded | 6996990909c33900fcd71efc | | Memobase | 75.78 | Not recorded | 69eaf94055a3c32af89f8b75 | | Zep | 75.14 | Not recorded | 69eaf94055a3c32af89f8b75 | | RAG-Utterance | 74.87 | Not recorded | 69647fe753fdd5c361744743 | | FullContext | 73.83 | Not recorded | 69eaf94055a3c32af89f8b75 | | MMA | 72.31 | Not recorded | 6996990909c33900fcd71efc | | OpenAI | 71.82 | Not recorded | 69eaf94055a3c32af89f8b75 | | RAG-Session | 70.35 | Not recorded | 69647fe753fdd5c361744743 | | LightRAG | 68.83 | Not recorded | 69eaf94055a3c32af89f8b75 | | Mem0g | 68.44 | Not recorded | 69eaf94055a3c32af89f8b75 | | MemBuilder | 68.07 | Not recorded | 69647fe753fdd5c361744743 | | Mem0 | 66.88 | Not recorded | 69eaf94055a3c32af89f8b75 | | A-Mem | 64.16 | Not recorded | 69eaf94055a3c32af89f8b75 | | MiniRAG | 63.51 | Not recorded | 69eaf94055a3c32af89f8b75 | | Memory-R1 | 62.74 | Not recorded | 69647fe753fdd5c361744743 | | MemoryOS | 58.25 | Not recorded | 69eaf94055a3c32af89f8b75 | | LangMem | 58.1 | Not recorded | 69eaf94055a3c32af89f8b75 | | Mem0 | 51.64 | Not recorded | 69647fe753fdd5c361744743 | ### MemBuilder Accuracy: 84.23. Result ID: 6996957d5698665ec4dedd8c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction model":"Qwen3-4B","Training stage":"SFT + RL","Answer model":"Claude 4.5 Sonnet","Memory construction method":"Training-based"}. This record does not assert independent reproduction. ### MemBuilder Accuracy: 82.61. Result ID: 6996957d5698665ec4dedd87. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction model":"Claude 4.5 Sonnet","Memory construction method":"Prompting-based","Answer model":"Claude 4.5 Sonnet"}. This record does not assert independent reproduction. ### MemBuilder Accuracy: 81.74. Result ID: 6996957d5698665ec4dedd8a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction model":"Qwen3-4B","Training stage":"SFT","Answer model":"Claude 4.5 Sonnet","Memory construction method":"Training-based"}. This record does not assert independent reproduction. ### MemBuilder Accuracy: 79.91. Result ID: 6996957d5698665ec4dedd85. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction model":"GPT-4.1","Memory construction method":"Prompting-based","Answer model":"Claude 4.5 Sonnet"}. This record does not assert independent reproduction. ### MemBuilder Accuracy: 79.31. Result ID: 6996957d5698665ec4dedd8b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction model":"Qwen3-4B","Training stage":"RL","Answer model":"Claude 4.5 Sonnet","Memory construction method":"Training-based"}. This record does not assert independent reproduction. ### MIRIX Accuracy: 77.48. Result ID: 6996957d5698665ec4dedd84. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction method":"Prompting-based","Answer model":"Claude 4.5 Sonnet"}. This record does not assert independent reproduction. ### MemBuilder Accuracy: 77.47. Result ID: 6996957d5698665ec4dedd86. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction model":"QwQ-32B","Memory construction method":"Prompting-based","Answer model":"Claude 4.5 Sonnet"}. This record does not assert independent reproduction. ### MIRIX Accuracy: 77.37. Result ID: 699af1965698665ec482b621. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"baseline\_status":"true"}. This record does not assert independent reproduction. ### StructMem Accuracy: 76.82. Result ID: 69eb2d3e47f34ce924d387cd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"Structural Memory","Build Tokens (In)":"1.501","Build Tokens (Out)":"0.436","Build Tokens (Sum)":"1.937","Calls":"1056","Time (s)":"22854"}. This record does not assert independent reproduction. ### MMA Accuracy: 75.94. Result ID: 699af1965698665ec482b623. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"variant":"st","components":"S + T","consensus\_module":"disabled"}. This record does not assert independent reproduction. ### Memobase Accuracy: 75.78. Result ID: 69eb2d3e47f34ce924d387cc. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"Structural Memory"}. This record does not assert independent reproduction. ### Zep Accuracy: 75.14. Result ID: 69eb2d3e47f34ce924d387cb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"Structural Memory"}. This record does not assert independent reproduction. ### RAG-Utterance Accuracy: 74.87. Result ID: 6996957d5698665ec4dedd82. [Reporting paper]() Recorded evaluation configuration: {"Memory construction method":"Retrieval-based","Answer model":"Claude 4.5 Sonnet","Retrieval granularity":"Utterance-level"}. This record does not assert independent reproduction. ### FullContext Accuracy: 73.83. Result ID: 69eb2d3e47f34ce924d387c3. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"RAG"}. This record does not assert independent reproduction. ### MMA Accuracy: 72.31. Result ID: 699af1965698665ec482b622. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"components":"S + T + Ccon","variant":"Full Model"}. This record does not assert independent reproduction. ### OpenAI Accuracy: 71.82. Result ID: 69eb2d3e47f34ce924d387c2. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"RAG"}. This record does not assert independent reproduction. ### RAG-Session Accuracy: 70.35. Result ID: 6996957d5698665ec4dedd81. [Reporting paper]() Recorded evaluation configuration: {"Memory construction method":"Retrieval-based","Answer model":"Claude 4.5 Sonnet","Retrieval granularity":"Session-level"}. This record does not assert independent reproduction. ### LightRAG Accuracy: 68.83. Result ID: 69eb2d3e47f34ce924d387c5. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"RAG","Build Tokens (In)":"10.014","Build Tokens (Out)":"1.916","Build Tokens (Sum)":"11.931","Calls":"13576","Time (s)":"60469"}. This record does not assert independent reproduction. ### Mem0g Accuracy: 68.44. Result ID: 69eb2d3e47f34ce924d387ca. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"Structural Memory","Build Tokens (In)":"33.512","Build Tokens (Out)":"2.313","Build Tokens (Sum)":"35.825","Calls":"53514","Time (s)":"115670"}. This record does not assert independent reproduction. ### MemBuilder Accuracy: 68.07. Result ID: 6996957d5698665ec4dedd89. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction model":"Qwen3-4B","Training stage":"Base","Answer model":"Claude 4.5 Sonnet","Memory construction method":"Training-based"}. This record does not assert independent reproduction. ### Mem0 Accuracy: 66.88. Result ID: 69eb2d3e47f34ce924d387c8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"Flat Memory","Build Tokens (In)":"10.958","Build Tokens (Out)":"1.239","Build Tokens (Sum)":"12.196","Calls":"9181","Time (s)":"30057"}. This record does not assert independent reproduction. ### A-Mem Accuracy: 64.16. Result ID: 69eb2d3e47f34ce924d387c7. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"Flat Memory","Build Tokens (In)":"9.126","Build Tokens (Out)":"2.368","Build Tokens (Sum)":"11.494","Calls":"11754","Time (s)":"60607"}. This record does not assert independent reproduction. ### MiniRAG Accuracy: 63.51. Result ID: 69eb2d3e47f34ce924d387c4. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"RAG","Build Tokens (In)":"9.022","Build Tokens (Out)":"1.081","Build Tokens (Sum)":"10.103","Calls":"2508","Time (s)":"2566"}. This record does not assert independent reproduction. ### Memory-R1 Accuracy: 62.74. Result ID: 6996957d5698665ec4dedd88. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction model":"Llama-3.1-8B-Instruct","Answer model":"Llama-3.1-8B-Instruct","Memory construction method":"Training-based"}. This record does not assert independent reproduction. ### MemoryOS Accuracy: 58.25. Result ID: 69eb2d3e47f34ce924d387c9. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"Structural Memory","Build Tokens (In)":"1.889","Build Tokens (Out)":"0.939","Build Tokens (Sum)":"2.868","Calls":"5534","Time (s)":"24220"}. This record does not assert independent reproduction. ### LangMem Accuracy: 58.1. Result ID: 69eb2d3e47f34ce924d387c6. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"gpt-4o-mini","Embedding Model":"text-embedding-3-small","Method Category":"Flat Memory","Build Tokens (In)":"9.873","Build Tokens (Out)":"1.192","Build Tokens (Sum)":"11.066","Calls":"5990","Time (s)":"26281"}. This record does not assert independent reproduction. ### Mem0 Accuracy: 51.64. Result ID: 6996957d5698665ec4dedd83. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Memory construction method":"Prompting-based","Answer model":"Claude 4.5 Sonnet"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.