Interactive Agent Task on ALFWORLD
1Effective Steps MultiplierAGORA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| AGORABackbone=qwen3.5-flash, Per-step LLM call=false2026.05 | 1 | 0.001 | |
| AGORABackbone=gpt-5-mini, Per-step LLM call=false2026.05 | 1.4 | 0.0105 | |
| AGORABackbone=gpt-4o-mini, Per-step LLM call=false2026.05 | 1.7 | 0.0057 | |
| HiAgentBackbone=qwen3.5-flash, Per-step LLM call=true2026.05 | 2.4 | 0.0005 | |
| AgentDietBackbone=gpt-5-mini, Per-step LLM call=true2026.05 | 2.6 | 0.0052 | |
| ObsMaskBackbone=gpt-5-mini, Per-step LLM call=false2026.05 | 2.8 | 0.0048 | |
| AgentDietBackbone=qwen3.5-flash, Per-step LLM call=true2026.05 | 2.8 | 0.0004 | |
| ObsMaskBackbone=qwen3.5-flash, Per-step LLM call=false2026.05 | 2.9 | 0.0004 | |
| ACONBackbone=qwen3.5-flash, Per-step LLM call=true2026.05 | 3.3 | 0.0003 | |
| ObsMaskBackbone=gpt-4o-mini, Per-step LLM call=false2026.05 | 3.6 | 0.0016 | |
| HiAgentBackbone=gpt-4o-mini, Per-step LLM call=true2026.05 | 3.9 | 0.0015 | |
| ACONBackbone=gpt-5-mini, Per-step LLM call=true2026.05 | 4 | 0.0034 | |
| HiAgentBackbone=gpt-5-mini, Per-step LLM call=true2026.05 | 4.2 | 0.0029 | |
| AgentDietBackbone=gpt-4o-mini, Per-step LLM call=true2026.05 | 7.8 | 0.0007 | |
| ACONBackbone=gpt-4o-mini, Per-step LLM call=true2026.05 | 8.2 | 0.0007 |