Embodied Task Execution on ALFWorld
93.28Success RateEmbodiSkill w/ GPT-5.2
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| EmbodiSkill w/ GPT-5.2Executor=Qwen3.5-27B, Skill Evolution Model=GPT-5.22026.05 | 93.28 | — | 95.83 | 96.77 | 73.91 | 95.24 | 100 | 100 | |
| EmbodiSkill w/ Gemini-3-flashExecutor=Qwen3.5-27B, Skill Evolution Model=Gemini-3-flash2026.05 | 87.31 | — | 95.83 | 93.55 | 69.57 | 95.24 | 83.33 | 82.35 | |
| EmbodiSkill w/ GPT-5.2Executor=Qwen2.5-14B-Instruct, Skill Evolution Model=GPT-5.22026.05 | 86.57 | — | 91.67 | 96.77 | 73.91 | 95.24 | 83.33 | 70.59 | |
| EmbodiSkill w/ Gemini-3-flashExecutor=Qwen2.5-14B-Instruct, Skill Evolution Model=Gemini-3-flash2026.05 | 85.82 | — | 79.17 | 93.55 | 73.91 | 90.48 | 100 | 76.47 | |
| Gemini-3-flashExecutor=Gemini-3-flash2026.05 | 82.09 | — | 91.67 | 83.87 | 65.22 | 85.71 | 83.33 | 82.35 | |
| G-MemoryExecutor=Qwen3.5-27B, Memory Mechanism=G-Memory2026.05 | 74.62 | — | 62.5 | 77.42 | 82.61 | 85.71 | 83.33 | 52.94 | |
| GPT-5.2Executor=GPT-5.22026.05 | 70.89 | — | 87.5 | 67.74 | 56.52 | 76.19 | 83.33 | 52.94 | |
| G-MemoryExecutor=Qwen2.5-14B-Instruct, Memory Mechanism=G-Memory2026.05 | 67.16 | — | 50 | 83.87 | 69.57 | 85.71 | 83.33 | 17.65 | |
| Mem0Executor=Qwen3.5-27B, Memory Mechanism=Mem02026.05 | 64.18 | — | 62.5 | 61.29 | 56.52 | 85.71 | 72.22 | 47.06 | |
| LangMemExecutor=Qwen3.5-27B, Memory Mechanism=LangMem2026.05 | 62.69 | — | 62.5 | 61.29 | 52.17 | 76.19 | 72.22 | 52.94 | |
| No memoryExecutor=Qwen3.5-27B, Memory Mechanism=None2026.05 | 61.19 | — | 66.67 | 51.61 | 65.22 | 76.19 | 72.22 | 35.29 | |
| No memoryExecutor=Qwen2.5-14B-Instruct, Memory Mechanism=None2026.05 | 46.27 | — | 33.33 | 51.61 | 26.09 | 85.71 | 72.22 | 5.88 | |
| LangMemExecutor=Qwen2.5-14B-Instruct, Memory Mechanism=LangMem2026.05 | 36.57 | — | 12.5 | 32.26 | 26.09 | 71.43 | 83.33 | 0 | |
| MiniMax-M2.1Intervention Mechanism=Rollback, Calibration=Uncalibrated2026.02 | 16.6 | 0.5 | — | — | — | — | — | — | |
| MiniMax-M2.1Intervention Mechanism=Append, Calibration=Calibrated2026.02 | 16.5 | 0.5 | — | — | — | — | — | — | |
| MiniMax-M2.1Intervention Mechanism=Baseline, Calibration=N/A2026.02 | 16.1 | 0.5 | — | — | — | — | — | — | |
| GLM-4.7Intervention Mechanism=Rollback, Calibration=Uncalibrated2026.02 | 15.8 | 1.1 | — | — | — | — | — | — | |
| MiniMax-M2.1Intervention Mechanism=Rollback, Calibration=Calibrated2026.02 | 15.8 | 0.5 | — | — | — | — | — | — | |
| GLM-4.7Intervention Mechanism=Append, Calibration=Uncalibrated2026.02 | 15.3 | 1.1 | — | — | — | — | — | — | |
| MiniMax-M2.1Intervention Mechanism=Append, Calibration=Uncalibrated2026.02 | 15.3 | 0.5 | — | — | — | — | — | — | |
| GLM-4.7Intervention Mechanism=Rollback, Calibration=Calibrated2026.02 | 15 | 1.1 | — | — | — | — | — | — | |
| GLM-4.7Intervention Mechanism=Baseline, Calibration=N/A2026.02 | 14.7 | 1.1 | — | — | — | — | — | — | |
| GLM-4.7Intervention Mechanism=Append, Calibration=Calibrated2026.02 | 13.4 | 1.1 | — | — | — | — | — | — | |
| Mem0Executor=Qwen2.5-14B-Instruct, Memory Mechanism=Mem02026.05 | 12.69 | — | 0 | 0 | 4.35 | 0 | 88.89 | 0 | |
| Qwen-3-8BIntervention Mechanism=Append, Calibration=Uncalibrated2026.02 | 8.6 | 2.8 | — | — | — | — | — | — | |
| Qwen-3-8BIntervention Mechanism=Rollback, Calibration=Uncalibrated2026.02 | 7.9 | 2.8 | — | — | — | — | — | — | |
| Qwen-3-8BIntervention Mechanism=Append, Calibration=Calibrated2026.02 | 7.8 | 2.8 | — | — | — | — | — | — | |
| Qwen-3-8BIntervention Mechanism=Rollback, Calibration=Calibrated2026.02 | 6.9 | 2.8 | — | — | — | — | — | — | |
| Qwen-3-8BIntervention Mechanism=Baseline, Calibration=N/A2026.02 | 5.8 | 2.8 | — | — | — | — | — | — |