MemoryBench Task on MemoryBench Long-Input-Long-Output
64.23Norm-ScoreUNO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| UNOModel=Qwen3-8B, Method Type=Training2026.02 | 64.23 | 17.74 | |
| UNO-SingleModel=Qwen3-8B, Method Type=Training2026.02 | 63.77 | 15.66 | |
| BaseModel=Qwen3-8B, Method Type=Baseline2026.02 | 63.41 | 14.87 | |
| SFT w/ UNO DataModel=Qwen3-8B, Method Type=Training2026.02 | 62.24 | 8.98 | |
| EmbeddingModel=Qwen3-8B, Method Type=RAG2026.02 | 62.04 | 7.58 | |
| DPO w/ UNO DataModel=Qwen3-8B, Method Type=Training2026.02 | 61.65 | 6.87 | |
| BM25Model=Qwen3-8B, Method Type=RAG2026.02 | 61.49 | 6.5 | |
| A-MemModel=Qwen3-8B, Method Type=Memory2026.02 | 60.55 | 2.22 | |
| ReMemModel=Qwen3-8B, Method Type=Memory2026.02 | 59.09 | -4.3 | |
| Mem0Model=Qwen3-8B, Method Type=Memory2026.02 | 58.29 | -5.7 | |
| MemOSModel=Qwen3-8B, Method Type=Memory2026.02 | 57.95 | -7.71 | |
| UNOModel=phi-4, Method Type=Training2026.02 | 57.84 | -13.62 | |
| UNO-SingleModel=phi-4, Method Type=Training2026.02 | 57.14 | -17.1 | |
| BaseModel=phi-4, Method Type=Baseline2026.02 | 57.07 | -17.36 | |
| SFT w/ UNO DataModel=phi-4, Method Type=Training2026.02 | 55.63 | -24.44 | |
| EmbeddingModel=phi-4, Method Type=RAG2026.02 | 54.43 | -28.63 | |
| MemOSModel=phi-4, Method Type=Memory2026.02 | 53.71 | -31.43 | |
| BM25Model=phi-4, Method Type=RAG2026.02 | 53.67 | -31.6 | |
| A-MemModel=phi-4, Method Type=Memory2026.02 | 53.22 | -32.94 | |
| DPO w/ UNO DataModel=phi-4, Method Type=Training2026.02 | 52.54 | -36.95 | |
| Mem0Model=phi-4, Method Type=Memory2026.02 | 52.23 | -37.53 | |
| ReMemModel=phi-4, Method Type=Memory2026.02 | 47.72 | -54.9 | |
| MemoryOSModel=Qwen3-8B, Method Type=Memory2026.02 | 45.96 | -59.89 | |
| MemoryOSModel=phi-4, Method Type=Memory2026.02 | 24.23 | -161.25 |