Learning the Graph Language of Molecules: Grammar Tokenization for Chemical LLMs

Large language models (LLMs) have been explored for integrating molecular graphs, but joint autoregressive modeling of text and molecules remains challenging. A key barrier is the lack of principled molecular graph tokenizers (MGTs): string representations like SMILES rely on brittle syntax and canonicalization, limiting robustness, flexibility, and interpretability. We introduce AlphaGrammar, a neural-symbolic MGT that rewrites molecular graphs as linear derivations over a formal graph grammar with theoretical guarantees. Training optimizes the parameters of a potential function for grammar construction, and inference amortizes the cost of parsing during tokenization. We then introduce Gramole, a multimodal LLM built on AlphaGrammar, which interleaves text and molecules as rule sequences without the need for modality-specific adapters. Gramole offers $100$% validity via type-constrained decoding, expressivity by sampling high-diversity molecular sets, transferability by improving FDA drugs' predicted properties on all six endpoints under an independent property predictor (vs. four for the strongest baseline), with edits that stay $1.9\times$ closer to the parent drug, controllability using natural language interfaces, and synthesizability by achieving $1.3\times$ drug and $3.1\times$ material retrosynthesis success rate of the best multimodal LLM baseline at matched Qwen3-8B backbone. AlphaGrammar autonomously acquires rules that recapitulate known functional groups, offering a layer of interpretability absent in prior MGTs. We benchmark Gramole across $8$ diverse tasks in the design-make-test-analysis (DMTA) cycle and probe AlphaGrammar's mechanism in $4$ quantitative and qualitative case studies.