BioMedGPT-Mol: Multi-task Learning for Molecular Understanding and Generation

TL;DR

BioMedGPT-Mol is a multi-task trained molecular language model that enhances understanding and generation of molecules, achieving state-of-the-art results in retrosynthetic planning and demonstrating the effectiveness of large-scale, multi-task fine-tuning.

Contribution

This work introduces BioMedGPT-Mol, a novel multi-task learning framework for molecular science that leverages large datasets and reasoning models to improve molecular understanding and synthesis planning.

Findings

Achieves state-of-the-art performance on RetroBench.

Effectively adapts general-purpose reasoning models to molecular tasks.

Demonstrates superior efficacy as an end-to-end retrosynthetic planner.

Abstract

Molecules play a crucial role in biomedical research and discovery, particularly in the field of small molecule drug development. Given the rapid advancements in large language models, especially the recent emergence of reasoning models, it is natural to explore how a general-purpose language model can be efficiently adapted for molecular science applications. In this work, we introduce BioMedGPT-Mol, a molecular language model designed to support molecular understanding and generation tasks. By curating and unifying existing public instruction datasets, we have assembled a large-scale, comprehensive, and high-quality training dataset. The model is then fine-tuned through a meticulously designed multi-task learning framework. On a consolidated benchmark derived from LlaSMol, TOMG-Bench, and MuMOInstruct, BioMedGPT-Mol achieves remarkable performance. Our experimental results demonstrate that a general-purpose reasoning model can be effectively and efficiently post-trained into a professional molecular language model through a well-structured multi-task curriculum. Leveraging these capabilities, we further apply the model to multi-step retrosynthetic planning, achieving state-of-the-art performance on RetroBench and demonstrating its superior efficacy as an end-to-end retrosynthetic planner. We anticipate that our approach can be extended to other biomedical scientific domains.