ChemHTS introduces a novel approach that combines multiple AI tools in a hierarchical structure to significantly improve language models' performance on complex chemistry tasks.

• Overcomes tool invocation errors that limit current approaches
• Enables effective collaboration among diverse chemical analysis tools
• Enhances performance for molecular design, reaction prediction, and property estimation
• Creates a hierarchical structure where specialized tools work together systematically

This research represents a significant advancement for pharmaceutical development by improving AI systems' ability to design molecules, predict reactions, and estimate properties—potentially accelerating drug discovery processes and reducing development costs.

ChemHTS: Hierarchical Tool Stacking for Enhancing Chemical Agents