BioMaze tests and enhances large language models' ability to reason through complex biological pathways—essential for predicting biological phenomena and designing experiments.

• Introduces a 5,100-problem dataset derived from real biological research
• Evaluates LLMs on multi-hop reasoning in biological contexts
• Demonstrates how LLMs can support hypothesis generation and experimental design in biology

This research bridges the gap between AI capabilities and biological pathway understanding, providing a framework to enhance LLMs for scientific discovery in complex biological systems.

BioMaze: Benchmarking and Enhancing Large Language Models for Biological Pathway Reasoning