Property Prediction on PolyBench (test)
0.98rPolyLM (ChemLLM-7B)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PolyLM (ChemLLM-7B)category=PolyBench, CoT=false2026.01 | 0.98 | 19.9 | |
| PolyLM (ChemLLM-7B)category=PolyBench, CoT=true2026.01 | 0.98 | 20.3 | |
| PolyLM (Qwen-2.5-14B)category=PolyBench, CoT=true2026.01 | 0.98 | 21.7 | |
| PolyLM (Phi-4-14B)category=PolyBench, CoT=false2026.01 | 0.98 | 16.8 | |
| PolyLM (Phi-4-14B)category=PolyBench, CoT=true2026.01 | 0.98 | 18.1 | |
| GPT-4ocategory=Closed-source2026.01 | 0.98 | 22.7 | |
| Sonnetcategory=Closed-source2026.01 | 0.98 | 20.5 | |
| PolyLM (Qwen-2.5-14B)category=PolyBench, CoT=false2026.01 | 0.97 | 37.5 | |
| GPT-5category=Closed-source2026.01 | 0.97 | 16.47 | |
| Qwen-2.5-14Bcategory=Off-the-shelf2026.01 | 0.93 | 48.8 | |
| PolyLM (Qwen-2.5-7B)category=PolyBench, CoT=false2026.01 | 0.93 | 39.5 | |
| Phi-4-14Bcategory=Off-the-shelf2026.01 | 0.91 | 46.2 | |
| PolyLM (Qwen-2.5-7B)category=PolyBench, CoT=true2026.01 | 0.91 | 47.1 | |
| 4o-Minicategory=Closed-source2026.01 | 0.81 | 55.1 | |
| Qwen-2.5-7Bcategory=Off-the-shelf2026.01 | 0.78 | 73.6 | |
| ChemLLM-7Bcategory=Domain aligned2026.01 | 0.76 | 78 | |
| Llamole-8Bcategory=Domain aligned2026.01 | 0.13 | — | |
| ether0-24Bcategory=Domain aligned2026.01 | 0.1 | — | |
| LlaSMol-8Bcategory=Domain aligned2026.01 | 0.07 | 181.7 |