Embodied Task Completion on ALFWorld
94Success RateReAct
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ReActBackbone=DeepSeek-R12026.03 | 94 | 90.3 | |
| DECENTMEMBackbone=GEMMA4-E4B, Framework=AutoGen2026.05 | 92.28 | — | |
| AutoAgentBackbone=DeepSeek-R12026.03 | 91 | 87.3 | |
| G-MemoryBackbone=GEMMA4-E4B, Framework=AutoGen2026.05 | 89.17 | — | |
| ChatDevBackbone=GEMMA4-E4B, Framework=AutoGen2026.05 | 88.99 | — | |
| MetaGPTBackbone=GEMMA4-E4B, Framework=AutoGen2026.05 | 86.27 | — | |
| G-MemoryBackbone=GEMMA4-E4B, Framework=DyLAN2026.05 | 84.49 | — | |
| EQUIMEMFramework=AutoGen, Memory=G-Memory2026.05 | 84.4 | — | |
| DECENTMEMBackbone=GEMMA4-E4B, Framework=DyLAN2026.05 | 84.21 | — | |
| AutoAgentBackbone=QwQ-32B2026.03 | 82.8 | 82 | |
| No memoryBackbone=GEMMA4-E4B, Framework=AutoGen2026.05 | 82.73 | — | |
| G-MemoryBackbone=GEMMA4-E4B, Framework=AgentNet2026.05 | 82.39 | — | |
| EQUIMEMFramework=DyLAN, Memory=Voyager2026.05 | 81.6 | — | |
| ChatDevBackbone=GEMMA4-E4B, Framework=AgentNet2026.05 | 81.59 | — | |
| DECENTMEMBackbone=GEMMA4-E4B, Framework=AgentNet2026.05 | 81.46 | — | |
| EQUIMEMFramework=AutoGen, Memory=AriGraph2026.05 | 80.5 | — | |
| MetaGPTBackbone=GEMMA4-E4B, Framework=DyLAN2026.05 | 80.49 | — | |
| DeepAgentBackbone=QwQ-32B2026.03 | 79.9 | 87.9 | |
| EQUIMEMFramework=AutoGen, Memory=MemBank2026.05 | 79.2 | — | |
| PPL FilterFramework=AutoGen, Memory=G-Memory2026.05 | 78.4 | — | |
| MetaGPTBackbone=GEMMA4-E4B, Framework=AgentNet2026.05 | 77.47 | — | |
| ChatDevBackbone=GEMMA4-E4B, Framework=DyLAN2026.05 | 77.23 | — | |
| LLM AuditFramework=AutoGen, Memory=G-Memory2026.05 | 76.6 | — | |
| EQUIMEMFramework=AutoGen, Memory=Voyager2026.05 | 76.5 | — | |
| PPL FilterFramework=AutoGen, Memory=AriGraph2026.05 | 76.5 | — | |
| EQUIMEMFramework=DyLAN, Memory=G-Memory2026.05 | 76.3 | — | |
| No memoryBackbone=GEMMA4-E4B, Framework=DyLAN2026.05 | 76.28 | — | |
| No memoryBackbone=GEMMA4-E4B, Framework=AgentNet2026.05 | 75.67 | — | |
| LLM AuditFramework=AutoGen, Memory=AriGraph2026.05 | 75.2 | — | |
| ChatDevBackbone=GEMMA4-E2B, Framework=AutoGen2026.05 | 74.59 | — | |
| A-MemGuardFramework=DyLAN, Memory=Voyager2026.05 | 74.3 | — | |
| EQUIMEMFramework=MacNet, Memory=G-Memory2026.05 | 74.2 | — | |
| A-MemGuardFramework=AutoGen, Memory=MemBank2026.05 | 74 | — | |
| DECENTMEMBackbone=GEMMA4-E2B, Framework=AutoGen2026.05 | 73.43 | — | |
| EQUIMEMFramework=MacNet, Memory=AriGraph2026.05 | 73.1 | — | |
| EQUIMEMFramework=DyLAN, Memory=AriGraph2026.05 | 73.1 | — | |
| A-MemGuardFramework=AutoGen, Memory=Voyager2026.05 | 72.7 | — | |
| PPL FilterFramework=MacNet, Memory=G-Memory2026.05 | 72.6 | — | |
| VanillaFramework=AutoGen, Memory=G-Memory2026.05 | 72.1 | — | |
| PPL FilterFramework=DyLAN, Memory=G-Memory2026.05 | 71.5 | — | |
| EQUIMEMFramework=DyLAN, Memory=MemBank2026.05 | 71.3 | — | |
| LLM AuditFramework=DyLAN, Memory=AriGraph2026.05 | 71.1 | — | |
| PPL FilterFramework=AutoGen, Memory=MemBank2026.05 | 70.9 | — | |
| DECENTMEMBackbone=GEMMA4-E2B, Framework=AgentNet2026.05 | 70.8 | — | |
| EQUIMEMFramework=MacNet, Memory=Voyager2026.05 | 70.4 | — | |
| DECENTMEMBackbone=GEMMA4-E2B, Framework=DyLAN2026.05 | 69.53 | — | |
| VanillaFramework=AutoGen, Memory=AriGraph2026.05 | 69.4 | — | |
| LLM AuditFramework=AutoGen, Memory=MemBank2026.05 | 68.6 | — | |
| LLM AuditFramework=MacNet, Memory=G-Memory2026.05 | 68.5 | — | |
| G-MemoryBackbone=GEMMA4-E2B, Framework=AgentNet2026.05 | 68.47 | — | |
| G-MemoryBackbone=GEMMA4-E2B, Framework=DyLAN2026.05 | 68.11 | — | |
| A-MemGuardFramework=DyLAN, Memory=MemBank2026.05 | 67.5 | — | |
| MetaGPTBackbone=GEMMA4-E2B, Framework=AutoGen2026.05 | 67.49 | — | |
| LLM AuditFramework=DyLAN, Memory=G-Memory2026.05 | 67.4 | — | |
| VanillaFramework=MacNet, Memory=G-Memory2026.05 | 67.1 | — | |
| G-MemoryBackbone=GEMMA4-E2B, Framework=AutoGen2026.05 | 66.73 | — | |
| LLM AuditFramework=AutoGen, Memory=Voyager2026.05 | 66.5 | — | |
| PPL FilterFramework=MacNet, Memory=AriGraph2026.05 | 66.4 | — | |
| No memoryBackbone=GEMMA4-E2B, Framework=AutoGen2026.05 | 66.38 | — | |
| PPL FilterFramework=DyLAN, Memory=AriGraph2026.05 | 66.2 | — | |
| LLM AuditFramework=DyLAN, Memory=Voyager2026.05 | 66.1 | — | |
| EQUIMEMFramework=MacNet, Memory=MemBank2026.05 | 66 | — | |
| A-MemGuardFramework=MacNet, Memory=Voyager2026.05 | 65.8 | — | |
| VanillaFramework=AutoGen, Memory=Voyager2026.05 | 65 | — | |
| VanillaFramework=AutoGen, Memory=MemBank2026.05 | 64.9 | — | |
| LLM AuditFramework=MacNet, Memory=AriGraph2026.05 | 64.8 | — | |
| PPL FilterFramework=DyLAN, Memory=Voyager2026.05 | 64.8 | — | |
| MetaGPTBackbone=GEMMA4-E2B, Framework=DyLAN2026.05 | 64.51 | — | |
| PPL FilterFramework=AutoGen, Memory=Voyager2026.05 | 64.2 | — | |
| VanillaFramework=DyLAN, Memory=AriGraph2026.05 | 64.1 | — | |
| MetaGPTBackbone=GEMMA4-E2B, Framework=AgentNet2026.05 | 63.37 | — | |
| ChatDevBackbone=GEMMA4-E2B, Framework=DyLAN2026.05 | 63 | — | |
| PPL FilterFramework=DyLAN, Memory=MemBank2026.05 | 62.8 | — | |
| No memoryBackbone=GEMMA4-E2B, Framework=DyLAN2026.05 | 62.29 | — | |
| ChatDevBackbone=GEMMA4-E2B, Framework=AgentNet2026.05 | 62.27 | — | |
| No-memoryFramework=AutoGen, Memory=Stateless2026.05 | 62.1 | — | |
| VanillaFramework=DyLAN, Memory=Voyager2026.05 | 61.9 | — | |
| No memoryBackbone=GEMMA4-E2B, Framework=AgentNet2026.05 | 60.43 | — | |
| LLM AuditFramework=MacNet, Memory=MemBank2026.05 | 60.3 | — | |
| PPL FilterFramework=MacNet, Memory=Voyager2026.05 | 60.1 | — | |
| LLM AuditFramework=DyLAN, Memory=MemBank2026.05 | 59.3 | — | |
| VanillaFramework=DyLAN, Memory=G-Memory2026.05 | 59.1 | — | |
| A-MemGuardFramework=MacNet, Memory=MemBank2026.05 | 58.5 | — | |
| VanillaFramework=MacNet, Memory=AriGraph2026.05 | 58.3 | — | |
| AutoAgentBackbone=Qwen3-30B-A3B2026.03 | 58.2 | 61.2 | |
| LLM AuditFramework=MacNet, Memory=Voyager2026.05 | 57.2 | — | |
| PPL FilterFramework=MacNet, Memory=MemBank2026.05 | 57.2 | — | |
| VanillaFramework=MacNet, Memory=Voyager2026.05 | 56.4 | — | |
| VanillaFramework=DyLAN, Memory=MemBank2026.05 | 55.2 | — | |
| No-memoryFramework=DyLAN, Memory=Stateless2026.05 | 53.8 | — | |
| VanillaFramework=MacNet, Memory=MemBank2026.05 | 48.9 | — | |
| DeepAgentBackbone=Qwen3-30B-A3B2026.03 | 48.5 | 76.6 | |
| No-memoryFramework=MacNet, Memory=Stateless2026.05 | 47.3 | — | |
| Long context2026.06 | 46.7 | — | |
| SimpleMem2026.06 | 43.3 | — | |
| LightMem2026.06 | 43.3 | — | |
| ReActBackbone=QwQ-32B2026.03 | 41.8 | 40.3 | |
| DCI-Lite2026.06 | 41.7 | — | |
| DeepAgentBackbone=DeepSeek-R12026.03 | 40.3 | 62.4 | |
| HippoRAG2026.06 | 40 | — |