Interactive Task Completion on ALFWorld
100Pick Success RateRewardFlow
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| RewardFlowType=RL Training, Backbone=Qwen2.5-(VL)-7B-Instruct2026.03 | 100 | 78.6 | 100 | 81.3 | 81.8 | 85 | 89.8 | — | — | |
| DPEPOModel Backbone=Qwen2.5-1.5B-Instruct2026.04 | 100 | 100 | 96.3 | 93.8 | 96 | 87.5 | 95.7 | 92.5 | 94.1 | |
| DPEPOModel Backbone=Qwen2.5-7B-Instruct2026.04 | 100 | 100 | 96.3 | 93.8 | 100 | 100 | 98.6 | 97.8 | 98.2 | |
| SimpleMem2026.05 | 100 | 25 | 9.5 | 53.3 | 39.1 | 11.8 | — | — | 48.7 | |
| MemRL2026.05 | 100 | 33.3 | 10 | 46.2 | 9.5 | 6.7 | — | — | 41.3 | |
| RLSDBackbone=Qwen2.5-7B-Instruct2026.06 | 100 | 87.5 | 92.3 | 58.8 | 80 | 65.2 | 82 | — | — | |
| Skill1Backbone=Qwen2.5-7B-Instruct2026.06 | 100 | 83.3 | 89.5 | 100 | 88.5 | 85 | 92.2 | — | — | |
| SPEARModel Backbone=Qwen2.5-7B-Instruct2026.04 | 99.9 | 85.1 | 95.6 | 96.4 | 89.9 | 95.1 | 94.7 | — | — | |
| SAPOBackbone=Qwen2.5-7B-Instruct2026.06 | 98.7 | 73.9 | 98.1 | 92.6 | 85 | 89.2 | 92.2 | — | — | |
| SkillRLBackbone=Qwen2.5-7B-Instruct2026.06 | 97.9 | 71.4 | 90 | 90 | 95.5 | 87.5 | 89.9 | — | — | |
| GiGPOModel Backbone=Qwen2.5-7B-Instruct2026.04 | 97.7 | 82.7 | 98.8 | 83.7 | 89.3 | 79.2 | 90.8 | 90.2 | 90.5 | |
| GiGPOBackbone=Qwen2.5-7B-Instruct2026.06 | 97.7 | 82.7 | 98.8 | 83.7 | 89.3 | 79.2 | 90.8 | — | — | |
| SDARBackbone=Qwen2.5-3B-Instruct2026.06 | 97.1 | 62.5 | 100 | 61.9 | 75 | 84.2 | 84.4 | — | — | |
| StepOPSDBackbone=Qwen2.5-3B-Instruct2026.06 | 97.1 | 66.7 | 87 | 79.1 | 78.9 | 95 | 83.6 | — | — | |
| Skill1Backbone=Qwen2.5-3B-Instruct2026.06 | 96.7 | 76.9 | 96.3 | 93.8 | 77.3 | 95 | 90.6 | — | — | |
| GiGPOBackbone=Qwen3-1.7B2026.06 | 96.7 | 69.3 | 88.9 | 62.5 | 59.1 | 85 | 79.7 | — | — | |
| SPEARModel Backbone=Qwen2.5-1.5B-Instruct2026.04 | 96.4 | 86.5 | 96.1 | 99 | 87.6 | 91.6 | 93.2 | — | — | |
| SAMPOBackbone=Qwen3-4B2026.02 | 96.3 | 88.49 | 93.65 | 92.42 | 92.7 | 88.35 | 92.72 | — | — | |
| UCOBBackbone=Qwen3-1.7B2026.06 | 96.2 | 85.7 | 95.5 | 84.2 | 81.5 | 88.8 | 89.1 | — | — | |
| GIGPOBackbone=Qwen3-4B2026.02 | 94.8 | 83.03 | 86.37 | 81.15 | 75.38 | 59.21 | 81.09 | — | — | |
| SDARBackbone=Qwen2.5-7B-Instruct2026.06 | 94.7 | 75 | 100 | 86.7 | 68.2 | 78.9 | 85.9 | — | — | |
| RewardFlowType=RL Training, Backbone=Qwen2.5-(VL)-3B-Instruct2026.03 | 94.6 | 70 | 90 | 36.4 | 70 | 70 | 78.9 | — | — | |
| GiGPOModel Backbone=Qwen2.5-1.5B-Instruct2026.04 | 94.4 | 67.5 | 94.8 | 94.4 | 79.8 | 76.4 | 86.7 | 83.2 | 85 | |
| GiGPOBackbone=Qwen2.5-3B-Instruct2026.06 | 94.3 | 72.7 | 93.8 | 100 | 64.3 | 89.5 | 85.9 | — | — | |
| Skill-SDBackbone=Qwen2.5-7B-Instruct2026.06 | 93.9 | 93.8 | 90.9 | 100 | 69.2 | 68.4 | 85.1 | — | — | |
| D2SkillBackbone=Qwen2.5-7B-Instruct2026.06 | 93.8 | 77.8 | 94.7 | 95 | 95.5 | 72 | 87.8 | — | — | |
| UCOBBackbone=Qwen2.5-7B-Instruct2026.06 | 93.8 | 90.9 | 94.1 | 95.8 | 92.9 | 89.5 | 93 | — | — | |
| UCOBBackbone=Qwen2.5-3B-Instruct2026.06 | 93.8 | 93.3 | 95.8 | 94.1 | 91 | 78.9 | 92.2 | — | — | |
| SLIMEvaluation with retrieved external skills=true2026.05 | 92.9 | 100 | 91.4 | 78.3 | 88.5 | 81.2 | — | — | 87.5 | |
| Gemini-2.5-ProModel Type=Closed-source, Result source=SkillRL [59]2026.05 | 92.8 | 63.3 | 62.1 | 69 | 26.6 | 58.7 | — | — | 60.3 | |
| GRPOType=RL Training, Backbone=Qwen2.5-(VL)-7B-Instruct2026.03 | 92.3 | 53.3 | 90.5 | 77.8 | 69.6 | 47.6 | 75 | — | — | |
| GRPOBackbone=Qwen2.5-7B-Instruct2026.06 | 91.2 | 87.5 | 96.2 | 81 | 65 | 57.9 | 81.2 | — | — | |
| GRPOBackbone=Qwen2.5-3B-Instruct2026.06 | 91.2 | 62.5 | 96.2 | 61.9 | 65 | 47.4 | 75 | — | — | |
| GRPOModel Backbone=Qwen2.5-7B-Instruct2026.04 | 90.8 | 66.1 | 89.3 | 74.7 | 72.5 | 77.6 | 77.6 | 77.3 | 77.5 | |
| GSPOBackbone=Qwen3-4B2026.02 | 90.36 | 79.31 | 90.71 | 75.45 | 77.95 | 48.95 | 78.61 | — | — | |
| RLOOType=RL Training, Backbone=Qwen2.5-(VL)-7B-Instruct2026.03 | 90 | 85.7 | 88 | 50 | 85.7 | 37.5 | 75 | — | — | |
| SkillCBackbone=Qwen2.5-7B-Instruct2026.06 | 88.5 | 78.6 | 91.2 | 94.7 | 100 | 88.9 | 90.6 | — | — | |
| Skill-SDBackbone=Qwen2.5-3B-Instruct2026.06 | 88.2 | 50 | 96.2 | 52.4 | 65 | 57.9 | 73.4 | — | — | |
| RLSDBackbone=Qwen2.5-3B-Instruct2026.06 | 87.9 | 75 | 90.9 | 75 | 73.1 | 68.4 | 79.7 | — | — | |
| GRPOBackbone=Qwen3-4B2026.02 | 87.41 | 62.65 | 46.42 | 72.28 | 58.89 | 38.37 | 72.61 | — | — | |
| GRPOModel Backbone=Qwen2.5-1.5B-Instruct2026.04 | 85.3 | 53.7 | 84.5 | 78.2 | 59.7 | 53.5 | 72.8 | 71.1 | 72 | |
| RLOO2026.05 | 85.3 | 100 | 50 | 75 | 52.2 | 38.9 | — | — | 65.3 | |
| DAPO_GIGPOBackbone=Qwen3-4B2026.02 | 85.04 | 55.26 | 65.35 | 58.98 | 56.52 | 26.57 | 60.55 | — | — | |
| Gemini-2.5-proType=Closed-source/Multi-agent2026.02 | 84.97 | 61.61 | 63.94 | 22.22 | 62.5 | 75.25 | 66.41 | — | — | |
| GiGPOType=RL Training, Backbone=Qwen2.5-(VL)-7B-Instruct2026.03 | 84.6 | 80 | 96 | 71.4 | 73.9 | 84 | 82.8 | — | — | |
| EMPGBackbone=Qwen3-4B2026.02 | 84.18 | 61.53 | 69.83 | 72.49 | 46.51 | 0.04 | 57.91 | — | — | |
| GiGPOType=RL Training, Backbone=Qwen2.5-(VL)-3B-Instruct2026.03 | 82.1 | 50 | 76.9 | 53.3 | 60.9 | 50 | 64.8 | — | — | |
| DAPO_GRPOBackbone=Qwen3-4B2026.02 | 81.28 | 37.57 | 53.97 | 40.16 | 51.28 | 6.43 | 49.58 | — | — | |
| DeepSeek-R1Model Backbone=Frontier LLMs2026.04 | 80 | 46.2 | 77.8 | 56.3 | 80 | 87.5 | 75 | 85.1 | 80.1 | |
| GRPOType=RL Training, Backbone=Qwen2.5-(VL)-3B-Instruct2026.03 | 79.8 | 57.1 | 75.8 | 21.4 | 31.6 | 36.4 | 56.2 | — | — | |
| RLOOType=RL Training, Backbone=Qwen2.5-(VL)-3B-Instruct2026.03 | 78.1 | 46.7 | 75 | 31.3 | 43.5 | 33.3 | 55.5 | — | — | |
| RewardFlowType=RL Training, Backbone=Qwen2.5-1.5B-Instruct2026.03 | 77.4 | 64.3 | 84 | 62.5 | 86.4 | 25 | 68.8 | — | — | |
| CISPOBackbone=Qwen3-4B2026.02 | 76.03 | 37.12 | 58.56 | 50.97 | 57.88 | 23.68 | 54.42 | — | — | |
| GPT-4oModel Backbone=Frontier LLMs2026.04 | 75.3 | 60.8 | 31.2 | 56.7 | 21.6 | 49.8 | 48 | 66 | 57 | |
| GPT-4oModel Type=Closed-source, Result source=SkillRL [59]2026.05 | 75.3 | 60.8 | 31.2 | 56.7 | 21.6 | 49.8 | — | — | 48 | |
| SDARBackbone=Qwen3-1.7B2026.06 | 73.5 | 25 | 76.9 | 33.3 | 40 | 36.8 | 53.9 | — | — | |
| DeepSeek-V3Model Backbone=Frontier LLMs2026.04 | 71.4 | 61.5 | 77.8 | 81.3 | 76 | 75 | 74.3 | 80.6 | 77.5 | |
| Skill1Backbone=Qwen3-1.7B2026.06 | 71.4 | 54.5 | 59.4 | 72.7 | 60 | 73.7 | 65.6 | — | — | |
| GRPOBackbone=Qwen3-1.7B2026.06 | 71.1 | 41.7 | 36.4 | 40 | 31.8 | 31.6 | 46.1 | — | — | |
| GPT-5.2Type=Closed-source/Multi-agent2026.02 | 70.03 | 66.07 | 35.37 | 62.3 | 52.08 | 37.36 | 51.56 | — | — | |
| DebateType=Closed-source/Multi-agent2026.02 | 67.74 | 64.28 | 33.33 | 60 | 52.38 | 65 | 56.25 | — | — | |
| StepOPSDBackbone=Qwen3-1.7B2026.06 | 64.7 | 44.4 | 56.5 | 60.9 | 42.1 | 55 | 56.3 | — | — | |
| GRPOType=RL Training, Backbone=Qwen2.5-1.5B-Instruct2026.03 | 62.9 | 53.3 | 50 | 40 | 45.8 | 38.1 | 50 | — | — | |
| GPT-4oType=Closed-source/Multi-agent2026.02 | 61.11 | 33.33 | 36.36 | 50 | 45.45 | 63.64 | 50 | — | — | |
| GiGPOType=RL Training, Backbone=Qwen2.5-1.5B-Instruct2026.03 | 59.4 | 46.7 | 62.5 | 44.4 | 43.5 | 56.3 | 53.1 | — | — | |
| RLOOType=RL Training, Backbone=Qwen2.5-1.5B-Instruct2026.03 | 56.2 | 46.7 | 62.5 | 50 | 43.5 | 27.8 | 49.2 | — | — | |
| Skill-SDBackbone=Qwen3-1.7B2026.06 | 52.9 | 37.5 | 69.2 | 42.9 | 60 | 36.8 | 52.3 | — | — | |
| RLSDBackbone=Qwen3-1.7B2026.06 | 50 | 37.5 | 61.5 | 19 | 50 | 21.1 | 42.2 | — | — | |
| BaseType=Prompting, Backbone=Qwen2.5-(VL)-3B-Instruct2026.03 | 36.4 | 42.9 | 9.1 | 7.1 | 5.3 | 4.5 | 16.4 | — | — | |
| SAPOBackbone=Qwen3-4B2026.02 | 34.49 | 32.19 | 24.13 | 24.92 | 16.21 | 9.37 | 25.16 | — | — | |
| BaseType=Prompting, Backbone=Qwen2.5-(VL)-7B-Instruct2026.03 | 33.3 | 13.3 | 10.7 | 0 | 4.3 | 0 | 10.9 | — | — | |
| BaseType=Prompting, Backbone=Qwen2.5-1.5B-Instruct2026.03 | 5.9 | 5.5 | 3.3 | 9.7 | 4.2 | 0 | 4.1 | — | — | |
| ETOModel Backbone=Qwen2.5-1.5B-Instruct2026.04 | — | — | — | — | — | — | 64.1 | 66.4 | 65.3 | |
| ETOModel Backbone=Qwen2.5-7B-Instruct2026.04 | — | — | — | — | — | — | 70.3 | 74.2 | 72.3 | |
| RLVMRModel Backbone=Qwen2.5-1.5B-Instruct2026.04 | — | — | — | — | — | — | 89.1 | 87.9 | 88.5 | |
| RLVMRModel Backbone=Qwen2.5-7B-Instruct2026.04 | — | — | — | — | — | — | 91.4 | 91.8 | 91.6 |