Molecule Description Generation on ChEBI-20 (test)
66.6BLEU-2BioT5+
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| BioT5+INPUT=SMILES, Category=Specialist2026.01 | 66.6 | 59.1 | — | 58.4 | 65 | 68.1 | |
| BioT5INPUT=SMILES, Category=Specialist2026.01 | 63.5 | 55.6 | — | 55.9 | 63.3 | 65.6 | |
| MolCAINPUT=SMILES, Category=Specialist2026.01 | 62 | 53.1 | — | 53.7 | 61.8 | 65.1 | |
| MolFMINPUT=SMILES, Category=Specialist2026.01 | 58.5 | 49.8 | — | 50.8 | 59.4 | 60.7 | |
| DeepMoLMINPUT=Image2026.01 | 55.76 | 46.29 | — | 48.07 | 55.96 | 58.23 | |
| Mol-VL-7BINPUT=Image, Parameters=7B2026.01 | 55.73 | 46.14 | — | 47.26 | 56.61 | 58.14 | |
| MoMuINPUT=SMILES, Category=Specialist2026.01 | 54.9 | 46.2 | — | 47.9 | 57.5 | 57.6 | |
| DoubleCheck & BioT5+INPUT=Image2026.01 | 54.4 | 45.56 | — | 47.21 | 55.94 | 58.07 | |
| MolScribe & BioT5+INPUT=Image2026.01 | 53.43 | 44.64 | — | 46.48 | 55.3 | 57.09 | |
| Mol-VL-2BINPUT=Image, Parameters=2B2026.01 | 50.57 | 40.26 | — | 41.99 | 52.25 | 53.34 | |
| GIT-MolINPUT=SMILES, Category=Generalist2026.01 | 35.2 | 26.3 | — | 48.5 | 56 | 43 | |
| BioMedGPTINPUT=SMILES, Category=Generalist2026.01 | 30.6 | 19.8 | — | 25.7 | 38.3 | 35 | |
| Mol-InstructionsINPUT=SMILES, Category=Generalist2026.01 | 24.9 | 17.1 | — | 20.3 | 28.9 | 27.1 | |
| 3D-MoLMINPUT=SMILES, Category=Generalist2026.01 | 6.7 | 3 | — | 4.2 | 8.6 | 18.3 | |
| Qwen2-VL-7BINPUT=Image, Parameters=7B2026.01 | 6.34 | 1.13 | — | 4.53 | 14.15 | 19.95 | |
| Text+Chem T5-augm-baseModel Category=Specialist Models2023.11 | 0.625 | 0.542 | 0.682 | 0.543 | 0.622 | 0.648 | |
| MoICAModel Category=Specialist Models, Backbone=Galac1.3B2023.11 | 0.62 | 0.531 | 0.681 | 0.537 | 0.618 | 0.651 | |
| GPT-4-0314Model Category=Retrieval Based LLMs, Setting=10-shot, Framework=MolReGPT2023.11 | 0.607 | 0.525 | 0.634 | 0.476 | 0.562 | 0.61 | |
| MolXPTModel Category=Specialist Models2023.11 | 0.594 | 0.505 | 0.66 | 0.511 | 0.597 | 0.626 | |
| MolFMModel Category=Specialist Models, Backbone=MolT5-base2023.11 | 0.585 | 0.498 | 0.653 | 0.508 | 0.594 | 0.607 | |
| GPT-3.5-turboModel Category=Retrieval Based LLMs, Setting=10-shot, Framework=MolReGPT2023.11 | 0.565 | 0.482 | 0.623 | 0.45 | 0.543 | 0.585 | |
| MoMuModel Category=Specialist Models, Backbone=MolT5-base2023.11 | 0.549 | 0.462 | — | — | 0.576 | 0.536 | |
| MolT5-baseModel Category=Specialist Models2023.11 | 0.54 | 0.457 | 0.634 | 0.485 | 0.568 | 0.569 | |
| InstructMol-GSModel Category=LLM Based Generalist Models, Modality=Structural and sequential information2023.11 | 0.475 | 0.371 | 0.566 | 0.394 | 0.502 | 0.509 | |
| InstructMol-GModel Category=LLM Based Generalist Models, Modality=Structural information only2023.11 | 0.466 | 0.365 | 0.547 | 0.365 | 0.479 | 0.491 | |
| GIT-Mol-(graph+SMILES)Model Category=Specialist Models, Input Modality=graph+SMILES2023.11 | 0.352 | 0.263 | 0.575 | 0.485 | 0.56 | 0.43 | |
| GIT-Mol-graphModel Category=Specialist Models, Input Modality=graph2023.11 | 0.29 | 0.21 | 0.54 | 0.445 | 0.512 | 0.491 | |
| GIT-Mol-SMILESModel Category=Specialist Models, Input Modality=SMILES2023.11 | 0.264 | 0.176 | 0.477 | 0.374 | 0.451 | 0.43 | |
| Mol-InstructionModel Category=LLM Based Generalist Models2023.11 | 0.249 | 0.171 | 0.331 | 0.203 | 0.289 | 0.271 | |
| BioMedGPT-10BModel Category=LLM Based Generalist Models2023.11 | 0.234 | 0.141 | 0.386 | 0.206 | 0.332 | 0.308 | |
| BioMedGPT-10BAlignment=w/2023.08 | 0.234 | 0.141 | 0.386 | 0.206 | 0.332 | 0.308 | |
| GPT-3.5-turboModel Category=LLM Based Generalist Models, Setting=zero-shot2023.11 | 0.103 | 0.05 | 0.261 | 0.088 | 0.204 | 0.161 | |
| ChatGPTAlignment=w/o2023.08 | 0.103 | 0.05 | 0.261 | 0.088 | 0.204 | 0.161 | |
| Llama2-7B-ChatAlignment=w/o2023.08 | 0.075 | 0.009 | 0.184 | 0.043 | 0.142 | 0.149 |