Multi-turn Planning on ALFWorld (test)
97.9RewardGOSKILLS
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GOSKILLSModel=Gemini 3 Pro, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=Yes2026.05 | 97.9 | 27,215 | 49.2 | |
| GOSKILLSModel=Claude Sonnet 4.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=Yes2026.05 | 96.4 | 58,904 | 39.5 | |
| GOSKILLSModel=GPT-5.4, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=Yes2026.05 | 95.3 | 63,319 | 38.2 | |
| Graph of SkillsModel=Gemini 3 Pro, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=dependency-graph2026.05 | 95 | 29,846 | 42.6 | |
| Graph of SkillsModel=Claude Sonnet 4.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=dependency-graph2026.05 | 94.6 | 49,120 | 62.7 | |
| Vector SkillsModel=Claude Sonnet 4.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=No2026.05 | 94.3 | 36,950 | 55.8 | |
| Vector SkillsModel=Gemini 3 Pro, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=No2026.05 | 93.6 | 28,407 | 37.8 | |
| Graph of SkillsModel=GPT-5.4, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=dependency-graph2026.05 | 93.6 | 47,851 | 65 | |
| Vector SkillsModel=GPT-5.4, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=No2026.05 | 92.9 | 34,436 | 57 | |
| GOSKILLSModel=Kimi K2.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=Yes2026.05 | 92.1 | 60,755 | 40.8 | |
| Vanilla SkillsModel=Claude Sonnet 4.5, Skill Access=Yes, Bounded Retrieval=No, Group Context=No2026.05 | 91.4 | 1,377,280 | 78.9 | |
| Graph of SkillsModel=Kimi K2.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=dependency-graph2026.05 | 90.7 | 54,336 | 63.1 | |
| Vanilla SkillsModel=Gemini 3 Pro, Skill Access=Yes, Bounded Retrieval=No, Group Context=No2026.05 | 89.3 | 1,524,401 | 53.2 | |
| Vanilla SkillsModel=GPT-5.4, Skill Access=Yes, Bounded Retrieval=No, Group Context=No2026.05 | 89.3 | 1,435,614 | 83.3 | |
| Vector SkillsModel=Kimi K2.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=No2026.05 | 89 | 41,908 | 58.4 | |
| No SkillsModel=GPT-5.4, Skill Access=No, Bounded Retrieval=No, Group Context=No2026.05 | 85.7 | 1,024,680 | 91.5 | |
| No SkillsModel=Claude Sonnet 4.5, Skill Access=No, Bounded Retrieval=No, Group Context=No2026.05 | 84.7 | 1,208,300 | 82.4 | |
| Vanilla SkillsModel=Kimi K2.5, Skill Access=Yes, Bounded Retrieval=No, Group Context=No2026.05 | 84.3 | 1,604,870 | 76.2 | |
| No SkillsModel=Gemini 3 Pro, Skill Access=No, Bounded Retrieval=No, Group Context=No2026.05 | 82.1 | 1,086,420 | 60.9 | |
| GOSKILLSModel=Qwen 3.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=Yes2026.05 | 80.7 | 74,502 | 55.3 | |
| No SkillsModel=Kimi K2.5, Skill Access=No, Bounded Retrieval=No, Group Context=No2026.05 | 78.6 | 1,156,200 | 84.9 | |
| Graph of SkillsModel=Qwen 3.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=dependency-graph2026.05 | 76.4 | 69,778 | 70.6 | |
| Vector SkillsModel=Qwen 3.5, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=No2026.05 | 72.9 | 58,214 | 76 | |
| Vanilla SkillsModel=Qwen 3.5, Skill Access=Yes, Bounded Retrieval=No, Group Context=No2026.05 | 69.3 | 1,742,510 | 91.4 | |
| No SkillsModel=Qwen 3.5, Skill Access=No, Bounded Retrieval=No, Group Context=No2026.05 | 61.4 | 1,318,760 | 98.1 | |
| GOSKILLSModel=MiniMax M2.7, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=Yes2026.05 | 54.3 | 65,227 | 68.8 | |
| Graph of SkillsModel=MiniMax M2.7, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=dependency-graph2026.05 | 52.1 | 67,884 | 71.5 | |
| Vector SkillsModel=MiniMax M2.7, Skill Access=Yes, Bounded Retrieval=Yes, Group Context=No2026.05 | 50.7 | 66,109 | 73.4 | |
| Vanilla SkillsModel=MiniMax M2.7, Skill Access=Yes, Bounded Retrieval=No, Group Context=No2026.05 | 47.1 | 2,184,823 | 88.6 | |
| No SkillsModel=MiniMax M2.7, Skill Access=No, Bounded Retrieval=No, Group Context=No2026.05 | 42 | 1,436,550 | 94 |