Long-term Memory Evaluation on LongMemEvalS
95.6Overall ScoreClaude Opus 4.6
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Claude Opus 4.62026.03 | 95.6 | 100 | 88.72 | 100 | 100 | 98.57 | 95.49 | |
| GPT-5-mini2026.03 | 94.2 | 96.15 | 85.71 | 98.21 | 100 | 97.14 | 96.99 | |
| Claude Sonnet 4.52026.03 | 94.2 | 98.72 | 89.47 | 100 | 93.33 | 98.57 | 91.73 | |
| Claude Haiku 4.52026.03 | 94 | 96.15 | 88.72 | 100 | 90 | 97.14 | 94.74 | |
| GPT-5.22026.03 | 93.8 | 97.44 | 83.46 | 100 | 100 | 98.57 | 95.49 | |
| MemMachineModel=gpt-5-mini, Ablation setting=Ablation best2026.04 | 93 | — | — | — | — | — | — | |
| GPT-4o2026.03 | 92.6 | 96.15 | 91.73 | 100 | 80 | 94.29 | 90.23 | |
| Claude Code Sonnet2026.03 | 88.6 | 93.59 | 79.7 | 94.64 | 80 | 94.29 | 90.98 | |
| MemCoTModel=GPT-4o-mini [4]2026.04 | 88 | 88.4 | 79.6 | 100 | 70 | 98.5 | 89.4 | |
| MnemisModel=GPT-4o-mini [4]2026.04 | 87.2 | 92.3 | 76.7 | 100 | 90 | 97.1 | 83.5 | |
| EMem-GModel=GPT-4o-mini [4]2026.04 | 77.9 | 94.4 | 73.6 | 87.5 | 32.2 | 87 | 74.8 | |
| Mem0Model=GPT-4o-mini [4]2026.04 | 71.1 | 74.4 | 66.2 | 96.4 | 34 | 91.4 | 63.9 | |
| RAGModel=GPT-4o-mini [4]2026.04 | 67.2 | 70.5 | 47.4 | 91.1 | 70 | 88.6 | 63.2 | |
| HeLa-MemBackbone=GPT-4o-mini, learning rate η=0.02, decay rate λ=0.995, spreading activation strength β=0.1, spreading threshold θ=0.4, keyword weight=0.7, max flipped items m=32026.04 | 65.4 | — | — | — | — | — | — | |
| NemoriModel=GPT-4o-mini [4]2026.04 | 64.2 | 61.5 | 51.1 | 83.9 | 46.7 | 88.6 | 61.7 | |
| ZepModel=GPT-4o-mini [4]2026.04 | 63.2 | 74.4 | 47.4 | 75 | 53.3 | 92.9 | 54.1 | |
| A-MEMBackbone=GPT-4o-mini2026.04 | 62.6 | — | — | — | — | — | — | |
| NaiveRAGBackbone=GPT-4o-mini2026.04 | 61 | — | — | — | — | — | — | |
| FullTextBackbone=GPT-4o-mini2026.04 | 56.8 | — | — | — | — | — | — | |
| Full ContextModel=GPT-4o-mini [4]2026.04 | 55 | 78.2 | 38.3 | 89.3 | 6.7 | 78.6 | 42.1 | |
| Mem0Backbone=GPT-4o-mini2026.04 | 53.61 | — | — | — | — | — | — | |
| MemoryOSBackbone=GPT-4o-mini2026.04 | 44.8 | — | — | — | — | — | — | |
| LangMemBackbone=GPT-4o-mini2026.04 | 37.2 | — | — | — | — | — | — |