# ToolMol: Evolutionary Agentic Framework for Multi-objective Drug Discovery

Source: [ToolMol: Evolutionary Agentic Framework for Multi-objective Drug Discovery](<https://www.sota2.com/research/paper/6a05571cdbc836b87d828d12>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-05-16T05:00:05.655Z

Authors: Andrew Y. Zhou, Sharvaree Vadgama, Sumanth Varambally, Peter Eckmann, Michael K. Gilson, Rose Yu

Published: 2026-05-12T21:58:14.000Z

## Abstract

Advances in large language models (LLMs) have recently opened new and promising avenues for small-molecule drug discovery. Yet existing LLM-based approaches for molecular generation often suffer from high rates of invalid and low-quality ligand candidates, a result of the syntactic limitations of current models with regard to molecular strings. In this paper, we introduce $\\texttt{ToolMol}$, an evolutionary agentic framework for de novo drug design. $\\texttt{ToolMol}$ combines a multi-objective genetic algorithm with an agentic LLM operator that iteratively updates the ligand population. We build a comprehensive toolbox of RDKit-backed functions that allows our agentic operator to consisently make precise ligand modifications. $\\texttt{ToolMol}$ achieves state-of-the-art performance on multi-objective property optimization tasks, discovering drug-like and synthesizable ligands that have $\>10\\%$ stronger predicted binding affinity compared to existing methods, evaluated on three protein targets. $\\texttt{ToolMol}$ ligands additionally achieve state-of-the-art results in gold-standard Absolute Binding Free Energy scores, gaining over existing methods by over $35\\%$. By studying chain-of-thought reasoning traces, we observe that tool-calling enables the model to more faithfully execute its planned modifications, efficiently exploiting the strong chemical prior knowledge in LLMs.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Multi-objective Drug Discovery on Aggregate c-MET, BRD4, ACAA1](<https://www.sota2.com/research/sota/multi-objective-drug-discovery-on-aggregate-c-met-brd4-acaa1.md>): Multi-objective Drug Discovery; Aggregate c-MET, BRD4, ACAA1; Average Rank: 1.56.

- [Molecular Generation on c-MET top 15 molecules](<https://www.sota2.com/research/sota/molecular-generation-on-c-met-top-15-molecules.md>): Molecular Generation; c-MET top 15 molecules; ABFE: 7.3.

- [Molecular Generation on BRD4 (top 15 molecules)](<https://www.sota2.com/research/sota/molecular-generation-on-brd4-top-15-molecules.md>): Molecular Generation; BRD4 (top 15 molecules); ABFE: 6.4.

- [Multi-objective Drug Discovery on c-MET MET\_HUMAN](<https://www.sota2.com/research/sota/multi-objective-drug-discovery-on-c-met-met_human.md>): Multi-objective Drug Discovery; c-MET MET\_HUMAN; Binding Affinity (BA): 11.

- [Multi-objective Drug Discovery on BRD4\_HUMAN](<https://www.sota2.com/research/sota/multi-objective-drug-discovery-on-brd4_human.md>): Multi-objective Drug Discovery; BRD4\_HUMAN; BA: 10.64.

- [Multi-objective Drug Discovery on ACAA1 THIK\_HUMAN](<https://www.sota2.com/research/sota/multi-objective-drug-discovery-on-acaa1-thik_human.md>): Multi-objective Drug Discovery; ACAA1 THIK\_HUMAN; Binding Affinity: 9.7.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2605.12784>)

[PDF](<https://arxiv.org/pdf/2605.12784.pdf>)