Multi-Property Molecular Optimization (DRD2+QED+plogP) on ZINC 250k
15Success Rate (SR)MolMem
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MolMemBackbone=Qwen2.5-1.5B, Optimization Paradigm=Expert2026.04 | 15 | 46 | 6.26 | |
| Direct RetrievalBackbone=N/A, Optimization Paradigm=Apprentice, Optimization Strategy=Retrieval-based2026.04 | 8 | 43 | 3.72 | |
| PEIT-LLMBackbone=Llama3.1-8B, Optimization Paradigm=Apprentice, Optimization Strategy=Task-Specific LLM2026.04 | 4 | 45 | 5.51 | |
| Reinvent 4Backbone=Transformer, Optimization Paradigm=Novice2026.04 | 3 | 51 | 4.28 | |
| Direct PromptBackbone=Qwen2.5-1.5B, Optimization Paradigm=Apprentice, Optimization Strategy=Prompting & SFT2026.04 | 2 | 54 | 4.6 | |
| SFT-onlyBackbone=Qwen2.5-3B, Optimization Paradigm=Apprentice, Optimization Strategy=Prompting & SFT2026.04 | 1.5 | 61 | 2.83 | |
| LlaSMolBackbone=Mistral-7B, Optimization Paradigm=Apprentice, Optimization Strategy=Task-Specific LLM2026.04 | 1.5 | 47 | 3.59 | |
| Direct PromptBackbone=Qwen2.5-3B, Optimization Paradigm=Apprentice, Optimization Strategy=Prompting & SFT2026.04 | 1 | 58 | 3.68 | |
| SFT-onlyBackbone=Qwen2.5-1.5B, Optimization Paradigm=Apprentice, Optimization Strategy=Prompting & SFT2026.04 | 1 | 61 | 2.61 | |
| Graph-GABackbone=N/A, Optimization Paradigm=Novice2026.04 | 0 | 51 | 3.79 | |
| QMOBackbone=GRU, Optimization Paradigm=Novice2026.04 | 0 | 52 | 1.14 | |
| ChemLLMBackbone=ChemLLM-7B, Optimization Paradigm=Apprentice, Optimization Strategy=Task-Specific LLM2026.04 | 0 | 52 | 2.85 | |
| GeLLM3OBackbone=Llama3.1-8B, Optimization Paradigm=Apprentice, Optimization Strategy=Task-Specific LLM2026.04 | 0 | 61 | 2.92 |