Science Experiment Execution on ScienceWorld (test)
51.51Success RateG-Mem
Evaluation Results
| Method | Links | |
|---|---|---|
| G-MemExecutor=GPT-5.22026.05 | 51.51 | |
| LangMemExecutor=Gemini-3-flash2026.05 | 50.82 | |
| A-MemExecutor=GPT-5.22026.05 | 50 | |
| MemCompilerExecutor=Qwen-3.5-27B2026.05 | 48.44 | |
| A-MemExecutor=Gemini-3-flash2026.05 | 47.78 | |
| LangMemExecutor=GPT-5.22026.05 | 46.67 | |
| MemCompilerExecutor=Qwen-2.5-14B2026.05 | 46.67 | |
| No MemExecutor=Gemini-3-flash2026.05 | 45.55 | |
| G-MemExecutor=Gemini-3-flash2026.05 | 44.44 | |
| Mem-0Executor=Gemini-3-flash2026.05 | 42.86 | |
| A-MemExecutor=Qwen-2.5-14B2026.05 | 34.44 | |
| G-MemExecutor=Qwen-2.5-14B2026.05 | 32.22 | |
| G-MemExecutor=Qwen-3.5-27B2026.05 | 32.22 | |
| A-MemExecutor=Qwen-3.5-27B2026.05 | 32.22 | |
| No MemExecutor=Qwen-2.5-14B2026.05 | 31.11 | |
| Mem-0Executor=Qwen-2.5-14B2026.05 | 27.78 | |
| Mem-0Executor=Qwen-3.5-27B2026.05 | 26.67 | |
| LangMemExecutor=Qwen-3.5-27B2026.05 | 26.67 | |
| No MemExecutor=GPT-5.22026.05 | 25.55 | |
| MemGenExecutor=Qwen-2.5-14B2026.05 | 24.65 | |
| Mem-0Executor=GPT-5.22026.05 | 22.22 | |
| No MemExecutor=Qwen-3.5-27B2026.05 | 21.11 | |
| LangMemExecutor=Qwen-2.5-14B2026.05 | 15.56 | |
| MemGenExecutor=Qwen-3.5-27B2026.05 | 14.32 |