This research demonstrates how large language models can mirror human expert thinking in chemistry when integrated with traditional search algorithms.

• Enables strategic synthesis planning and reaction mechanism elucidation
• Bridges the gap between specialized ML algorithms and human-like chemical reasoning
• Creates a new paradigm for computer-aided chemistry that mimics expert thought processes
• Offers potential breakthroughs in pharmaceutical development and drug discovery

For the medical industry, this approach could significantly accelerate drug discovery by providing AI-assisted synthesis planning that follows expert reasoning patterns rather than rigid algorithms.

Chemical reasoning in LLMs unlocks steerable synthesis planning and reaction mechanism elucidation