Bolek: A Multimodal Language Model for Molecular Reasoning

ArXi:2605.02745v1 Announce Type: new Molecular property models increasingly high-stakes drug-discovery decisions, but their outputs are often difficult to audit: classical predictors return scores without rationale, while language models can produce fluent explanations weakly grounded in the input molecule. Across these tasks, Bolek outperforms its Qwen3-4B-Instruct base on all endpoints in yes/no mode and on 13 of 15 in chain-of-thought mode, raising mean ROC/PR AUC from 0.55 to 0.76.