Memory-Augmented Agent Evaluation on MemoryAgentBench
28.7Single-Hop AccuracyMemOS + MemGate
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MemOS + MemGateBase Model=GPT-4o-mini2026.06 | 28.7 | 36.4 | 40.3 | 41.8 | 36.8 | |
| MemOSBase Model=GPT-4o-mini2026.06 | 27.8 | 35.3 | 39.2 | 40.7 | 35.8 | |
| Mem0 + MemGateBase Model=GPT-4o-mini2026.06 | 26.4 | 33.1 | 37.3 | 38.7 | 33.9 | |
| MemOS + MemGateBase Model=Qwen-3-8B2026.06 | 25.3 | 31.4 | 36.1 | 36.8 | 32.4 | |
| OpenClaw + MemGateBase Model=GPT-4o-mini2026.06 | 25.1 | 31.8 | 36.2 | 37.6 | 32.7 | |
| Mem0Base Model=GPT-4o-mini2026.06 | 25 | 32 | 36 | 37.5 | 32.6 | |
| OpenClawBase Model=GPT-4o-mini2026.06 | 24.2 | 30.7 | 35.4 | 36.1 | 31.6 | |
| MemOSBase Model=Qwen-3-8B2026.06 | 24.1 | 30.3 | 34.8 | 35.6 | 31.2 | |
| Mem0 + MemGateBase Model=Qwen-3-8B2026.06 | 23.6 | 28.4 | 33.1 | 34.2 | 29.7 | |
| A-Mem + MemGateBase Model=GPT-4o-mini2026.06 | 23.4 | 29.7 | 34.6 | 36.2 | 31 | |
| A-MemBase Model=GPT-4o-mini2026.06 | 22.7 | 28.2 | 33.1 | 35.4 | 29.9 | |
| Mem0Base Model=Qwen-3-8B2026.06 | 22.4 | 27.6 | 32.3 | 33.1 | 28.8 | |
| OpenClaw + MemGateBase Model=Qwen-3-8B2026.06 | 21.8 | 26.9 | 31.7 | 32.9 | 28.3 | |
| OpenClawBase Model=Qwen-3-8B2026.06 | 20.7 | 26.1 | 31.2 | 31.8 | 27.5 | |
| A-Mem + MemGateBase Model=Qwen-3-8B2026.06 | 20.2 | 25.7 | 30.8 | 31.6 | 27.1 | |
| A-MemBase Model=Qwen-3-8B2026.06 | 18.8 | 24.3 | 29.1 | 30.4 | 25.7 |