Embodied AI Planning on TDW-COOK
100Success RateParallel
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ParallelModel Scale=GPT-5, Execution Strategy=Parallel2026.04 | 100 | 14.72 | 6.2 | 47.6 | |
| AgenticCache (Ours+)Model Scale=GPT-5, Execution Strategy=AgenticCache (Ours+)2026.04 | 100 | 1.75 | 675 | 4.4 | |
| AgenticCache (Ours+)Model Scale=GPT-5-mini, Execution Strategy=AgenticCache (Ours+)2026.04 | 100 | 1.4 | 826 | 1 | |
| BaselineModel Scale=GPT-5, Execution Strategy=Baseline2026.04 | 94.44 | 12.86 | 3.3 | 21 | |
| ParallelModel Scale=GPT-5-mini, Execution Strategy=Parallel2026.04 | 94.44 | 6.5 | 4.9 | 6.5 | |
| SpeculativeModel Scale=GPT-5, Execution Strategy=Speculative2026.04 | 83.33 | 6.12 | 7.8 | 21.6 | |
| BaselineModel Scale=GPT-5-mini, Execution Strategy=Baseline2026.04 | 83.33 | 6.88 | 2.8 | 3.2 | |
| AgenticCache (Ours+)Model Scale=GPT-5-nano, Execution Strategy=AgenticCache (Ours+)2026.04 | 72.22 | 2.76 | 2 | 0.7 | |
| BaselineModel Scale=GPT-5-nano, Execution Strategy=Baseline2026.04 | 61.11 | 12.05 | 6.8 | 2.1 | |
| ParallelModel Scale=GPT-5-nano, Execution Strategy=Parallel2026.04 | 55.56 | 9.96 | 13.7 | 4.4 | |
| SpeculativeModel Scale=GPT-5-mini, Execution Strategy=Speculative2026.04 | 50 | 5.42 | 7.3 | 4.4 |