Tool-use agentic performance on AppWorld
64.88Avg@4SkillX
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SkillXModel=GLM-4.6, Experience extraction model alignment=Aligned with inference model2026.04 | 64.88 | 88.69 | |
| ExpeLModel=GLM-4.6, Experience extraction model alignment=Aligned with inference model2026.04 | 64.14 | 85.12 | |
| AWMModel=GLM-4.6, Experience extraction model alignment=Aligned with inference model2026.04 | 62.2 | 84.52 | |
| A-MemModel=GLM-4.6, Experience extraction model alignment=Aligned with inference model2026.04 | 60.57 | 83.93 | |
| No MemoryModel=GLM-4.6, Experience extraction model alignment=Aligned with inference model2026.04 | 60.27 | 83.33 | |
| SkillXModel=Kimi-K2-Instruct-0905, Experience extraction model alignment=GLM-4.6 used for experience extraction2026.04 | 56.4 | 81.55 | |
| ExpeLModel=Kimi-K2-Instruct-0905, Experience extraction model alignment=GLM-4.6 used for experience extraction2026.04 | 52.98 | 78.87 | |
| ExpeLModel=Kimi-K2-Instruct-0905, Experience extraction model alignment=Aligned with inference model2026.04 | 52.53 | 78.57 | |
| AWMModel=Kimi-K2-Instruct-0905, Experience extraction model alignment=GLM-4.6 used for experience extraction2026.04 | 50.6 | 76.49 | |
| AWMModel=Kimi-K2-Instruct-0905, Experience extraction model alignment=Aligned with inference model2026.04 | 49.7 | 76.19 | |
| No MemoryModel=Kimi-K2-Instruct-0905, Experience extraction model alignment=Aligned with inference model2026.04 | 46.88 | 70.24 | |
| A-MemModel=Kimi-K2-Instruct-0905, Experience extraction model alignment=Aligned with inference model2026.04 | 46.58 | 72.62 | |
| SkillXModel=Qwen3-32B, Experience extraction model alignment=GLM-4.6 used for experience extraction2026.04 | 35.12 | 58.93 | |
| AWMModel=Qwen3-32B, Experience extraction model alignment=GLM-4.6 used for experience extraction2026.04 | 34.45 | 56.25 | |
| ExpeLModel=Qwen3-32B, Experience extraction model alignment=GLM-4.6 used for experience extraction2026.04 | 32.94 | 58.78 | |
| ExpeLModel=Qwen3-32B, Experience extraction model alignment=Aligned with inference model2026.04 | 32.87 | 58.93 | |
| AWMModel=Qwen3-32B, Experience extraction model alignment=Aligned with inference model2026.04 | 30.8 | 55.95 | |
| No MemoryModel=Qwen3-32B, Experience extraction model alignment=Aligned with inference model2026.04 | 27.68 | 47.62 | |
| A-MemModel=Qwen3-32B, Experience extraction model alignment=Aligned with inference model2026.04 | 26.79 | 50.59 |