Paper 2510.16880
Chem-R: Learning to Reason as a Chemist
- Published
- Oct 2025
- Research lab
- Independent
- Citations
- 11
- GitHub
- 30 stars
01 In brief
Summary
Chem-R is a chemical reasoning model designed to emulate the deliberative processes of chemists, addressing three key challenges in current LLMs: lack of chemical fundamentals, unreliable reasoning trajectories, and imbalanced performance across tasks.
The model is trained via a three-phase framework: (1) Chemical Foundation Training, which fine-tunes on large-scale non-reasoning corpora to establish core chemical knowledge; (2) Chemical Reasoning Protocol Distillation, which extracts structured, expert-like reasoning protocols from a teacher model and uses them to synthesize high-quality CoT data for fine-tuning; and (3) Multi-task Group Relative Policy Optimization (GRPO), which uses a curriculum-like weighting scheme to balance performance across diverse tasks.
Evaluated on four benchmarks (ChemLLMBench, ChEBI-20, TOMG-Bench, USPTO) covering 9 macro-tasks and 25 sub-tasks, Chem-R achieves state-of-the-art results, surpassing models like Gemini-2.5-Pro and DeepSeek-R1 by up to 32% on molecular tasks and 48% on reaction tasks.
Ablation studies confirm the necessity of all three phases, and human expert evaluation shows Chem-R's reasoning chains are more chemically sound, logically coherent, and aligned with expert thinking.
The model also demonstrates strong out-of-distribution generalization, improving success rates on tasks like Solubility from 10% to 83%.
02 From the paper
Abstract
Although large language models (LLMs) have significant potential to advance chemical discovery, current LLMs lack core chemical knowledge, produce unreliable reasoning trajectories, and exhibit suboptimal performance across diverse chemical tasks. To address these challenges, we propose Chem-R, a generalizable Chemical Reasoning model designed to emulate the deliberative processes of chemists. Chem-R is trained through a three-phase framework that progressively builds advanced reasoning capabilities, including: 1) Chemical Foundation Training, which establishes core chemical knowledge. 2) Chemical Reasoning Protocol Distillation, incorporating structured, expert-like reasoning traces to guide systematic and reliable problem solving. 3) Multi-task Group Relative Policy Optimization that optimizes the model for balanced performance across diverse molecular- and reaction-level tasks. This structured pipeline enables Chem-R to achieve state-of-the-art performance on comprehensive benchmarks, surpassing leading large language models, including Gemini-2.5-Pro and DeepSeek-R1, by up to 32% on molecular tasks and 48% on reaction tasks. Meanwhile, Chem-R also consistently outperforms the existing chemical foundation models across both molecular and reaction level tasks. These results highlight Chem-R's robust generalization, interpretability, and potential as a foundation for next-generation AI-driven chemical discovery. The code and model are available at https://github.com/davidweidawang/Chem-R.