# NovoMolGen: Rethinking Molecular Language Model Pretraining

arXiv: https://arxiv.org/abs/2508.13408 · HTML: https://mlsanity.com/paper/2508.13408/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Clinical and biomedical applications, one of 17 directions in LLM reasoning.
- Builds on 56 of its 85 references that are in the map.
- First posted on arXiv 19 Aug 2025.

## Placed in

- LLM reasoning / [Clinical and biomedical applications](https://mlsanity.com/lr/clinical-and-biomedical-applications/)

## Builds on (most cited first, 30 of 56)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Semi-Supervised Classification with Graph Convolutional Networks](https://mlsanity.com/paper/1609.02907/) (2016)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Deep Unsupervised Learning using Nonequilibrium Thermodynamics](https://mlsanity.com/paper/1503.03585/) (2015)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [How Powerful are Graph Neural Networks?](https://mlsanity.com/paper/1810.00826/) (2018)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Neural Message Passing for Quantum Chemistry](https://mlsanity.com/paper/1704.01212/) (2017)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [Neural Machine Translation of Rare Words with Subword Units](https://mlsanity.com/paper/1508.07909/) (2015)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks](https://mlsanity.com/paper/1810.02244/) (2018)
- [Probing Classifiers: Promises, Shortcomings, and Advances](https://mlsanity.com/paper/2102.12452/) (2021)
- [Automatic chemical design using a data-driven continuous representation of molecules](https://mlsanity.com/paper/1610.02415/) (2016)
- [Equivariant Diffusion for Molecule Generation in 3D](https://mlsanity.com/paper/2203.17003/) (2022)
- [Junction Tree Variational Autoencoder for Molecular Graph Generation](https://mlsanity.com/paper/1802.04364/) (2018)
- [Subword Regularization: Improving Neural Network Translation Models with Multiple Subword Candidates](https://mlsanity.com/paper/1804.10959/) (2018)
- [OLMo: Accelerating the Science of Language Models](https://mlsanity.com/paper/2402.00838/) (2024)
- [MolGAN: An implicit generative model for small molecular graphs](https://mlsanity.com/paper/1805.11973/) (2018)
- [Provably Powerful Graph Networks](https://mlsanity.com/paper/1905.11136/) (2019)
- [GraphVAE: Towards Generation of Small Graphs Using Variational Autoencoders](https://mlsanity.com/paper/1802.03480/) (2018)
- [Score-based Generative Modeling of Graphs via the System of Stochastic Differential Equations](https://mlsanity.com/paper/2202.02514/) (2022)
- [Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Models](https://mlsanity.com/paper/1811.12823/) (2018)
- [Molecular De Novo Design through Deep Reinforcement Learning](https://mlsanity.com/paper/1704.07555/) (2017)
- [Molecular Transformer - A Model for Uncertainty-Calibrated Chemical Reaction Prediction](https://mlsanity.com/paper/1811.02633/) (2018)
- [Hierarchical Generation of Molecular Graphs using Structural Motifs](https://mlsanity.com/paper/2002.03230/) (2020)
- [ChemRL-GEM: Geometry Enhanced Molecular Representation Learning for Property Prediction](https://mlsanity.com/paper/2106.06130/) (2021)
- [Fr'echet ChemNet Distance: A metric for generative models for molecules in drug discovery](https://mlsanity.com/paper/1803.09518/) (2018)

## Abstract (arXiv)

Chemical Language Models (CLMs) are increasingly used in de novo drug design, driven by recent growth in model scale, compute, and dataset size. However, the relationship between design choices, training dynamics, and downstream generation quality remains poorly understood. We present a compute-controlled scaling study of CLMs comprising more than 30,000 experiments across molecular representations (SMILES, SELFIES, SAFE), tokenizations (atom-level and byte-pair encoding), model scales (0.5M-1B parameters), leakage-controlled datasets (MOSES, ChEMBL, PubChem, ZINC-22), and architectures (decoder-only and encoder-decoder). By fitting IsoFLOP profiles, we establish clear scaling trends in pretraining loss, but find that these improvements do not translate into comparable gains in goal-directed molecular design. Layer-wise probing and sparse autoencoder analysis reveal continued development of chemical representations: chemical syntax saturates early, while semantic properties emerge more slowly and become increasingly accessible with further training and model scale. These representational gains coexist with diminishing improvements in goal-directed generation under the evaluated protocols and oracle budgets. Our resulting suite of models, NovoMolGen, achieves state-of-the-art results, outperforming prior CLMs and specialized generative models in goal-directed molecular generation across drug discovery tasks. These findings expose a disconnect between chemical representation learning and downstream molecular design, motivating the development of pretraining paradigms that more directly learn chemical semantics.