Evaluating the Progression of Large Language Model Capabilities for Small-Molecule Drug Design
For researchers in drug discovery, this work provides a benchmark and shows that RL post-training can close capability gaps in LLMs for drug design.
The paper introduces a suite of chemically-grounded tasks for evaluating LLMs in small-molecule drug design, finding that frontier models are increasingly proficient but have significant room for improvement, especially with low data. RL-based post-training substantially improves performance, making a smaller model competitive with frontier models.
Large Language Models (LLMs) have the potential to accelerate small molecule drug design due to their ability to reason about information from diverse sources and formats. However, their practical utility remains unclear due to the lack of benchmarks that reflect real-world scenarios. In this work, we introduce a suite of chemically-grounded tasks spanning molecular property prediction, molecular representation transformations, and molecular design. Importantly, we formulate these tasks as reinforcement learning (RL) environments, enabling a unified approach for evaluation and post-training. Across three model families, we find that frontier models are increasingly proficient at chemical tasks, but that there is significant room for improvement, especially in experimental settings with low data. Critically, we show that RL-based post-training can substantially improve performance. A smaller model post-trained on our environments becomes competitive with state-of-the-art frontier models, despite a significantly weaker base model. This suggests a practical route toward employing LLMs in drug discovery; by combining carefully-designed evaluation tasks with targeted post-training, we can both elucidate and close critical capability gaps.