Dialogue Question Answering on LoCoMo 10 conversations (full release)
35.2MultiHop F1EVOLVEMEM
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| EVOLVEMEMBackbone=GPT-5.12026.05 | 35.2 | 33.9 | 33.7 | 35 | 54.1 | 38.7 | 51.9 | 51 | 90.2 | 87 | 57.2 | 53.6 | |
| SimpleMemBackbone=GPT-4o2026.05 | 31.8 | 32.8 | 19.5 | 39.3 | 23.5 | 20.6 | 40.2 | 34.7 | 80.2 | 69.1 | 43.2 | 40.7 | |
| EVOLVEMEMBackbone=GPT-4o2026.05 | 31.6 | 41 | 32.9 | 39.9 | 38.4 | 58.9 | 49.6 | 49 | 93.6 | 84.7 | 54.3 | 56.9 | |
| Mem0Backbone=GPT-4o2026.05 | 30.9 | 26.1 | 15.6 | 13.2 | 21.7 | 18.6 | 29.5 | 26.2 | 85.7 | 85.9 | 39.7 | 37.1 | |
| MemGPTBackbone=GPT-4o2026.05 | 30.5 | 25.7 | 18.8 | 16.6 | 24.6 | 22.7 | 30.5 | 26.8 | 84.3 | 83.9 | 40.4 | 37.6 | |
| SimpleMemBackbone=GPT-5.12026.05 | 30.5 | 29.4 | 17.8 | 16.5 | 27.2 | 25.3 | 30.5 | 27.6 | 80.7 | 80.2 | 41.8 | 37.6 | |
| A-MEMBackbone=GPT-4o2026.05 | 29.5 | 25 | 17.4 | 15.7 | 20 | 17.1 | 26.6 | 23.4 | 89.8 | 89.7 | 39.4 | 37 | |
| Claude-MemBackbone=GPT-4o2026.05 | 29.4 | 25.4 | 15.3 | 13.3 | 16.7 | 13.2 | 24.3 | 21.5 | 91.5 | 91.5 | 38.3 | 36 | |
| Mem0Backbone=GPT-5.12026.05 | 29.2 | 24.9 | 16 | 13.6 | 26.1 | 24.2 | 29.8 | 27.4 | 81.9 | 81.8 | 39 | 36.9 | |
| Claude-MemBackbone=GPT-5.12026.05 | 28.9 | 24.6 | 17.1 | 14.6 | 26.4 | 24.4 | 29.2 | 27.1 | 81.4 | 81.4 | 38.8 | 36.8 | |
| MemGPTBackbone=GPT-5.12026.05 | 28.8 | 24.3 | 16.5 | 14 | 24.9 | 23 | 29.4 | 27.3 | 80.6 | 80.3 | 38.5 | 36.4 | |
| MemVerseBackbone=GPT-5.12026.05 | 28.7 | 24 | 17.3 | 15 | 27.7 | 24.9 | 29.7 | 27.6 | 78 | 77.8 | 38.3 | 36.2 | |
| A-MEMBackbone=GPT-5.12026.05 | 28.7 | 24.2 | 16.4 | 14.4 | 24.6 | 21.8 | 28.4 | 26.3 | 82.6 | 82.5 | 38.5 | 36.5 | |
| MemVerseBackbone=GPT-4o2026.05 | 26 | 21.7 | 15.7 | 13 | 19.6 | 17.5 | 19.2 | 16.5 | 94.4 | 94.4 | 36.5 | 34.2 |