Decision Making on ALFWorld Avg v1
59.88Success Rate (SR)BeliefMem
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BeliefMemBackbone=Qwen3-Next-80B-A3B-Instruct, Memory Corpus Size=50%2026.05 | 59.88 | 29.56 | |
| BeliefMemBackbone=Qwen3-Next-80B-A3B-Instruct, Memory Corpus Size=Full2026.05 | 58.66 | 28.99 | |
| ReadAgentBackbone=Qwen3-Next-80B-A3B-Instruct2026.05 | 54.03 | 27.65 | |
| Mem0Backbone=Qwen3-Next-80B-A3B-Instruct2026.05 | 39.81 | 33.4 | |
| MemoryBankBackbone=Qwen3-Next-80B-A3B-Instruct2026.05 | 37.96 | 35.07 | |
| LangMemBackbone=Qwen3-Next-80B-A3B-Instruct2026.05 | 34.24 | 35.8 | |
| A-MEMBackbone=Qwen3-Next-80B-A3B-Instruct2026.05 | 27.1 | 39.66 | |
| No-MemoryBackbone=Qwen3-Next-80B-A3B-Instruct2026.05 | 22.35 | 40.92 | |
| MemoryOSBackbone=Qwen3-Next-80B-A3B-Instruct2026.05 | 18.98 | 42.69 |