LiteFold Launches LiteMol-1, a Multi-Molecule Diffusion Foundation Model for Drug Design
anshulkundaje · x · 2026-08-21
LiteFold introduces LiteMol-1, its first foundation model: a multi-molecule diffusion language model pre-trained from scratch.
Structure-based design models (BoltzGen, O-Design, RFdiffusion) are now the de facto standard for biomolecule design, but they are expensive at scale — campaigns often generate tens of thousands of designs and filter down to a handful. They are also optimized mainly around binding, ignoring ADME, toxicity, drug-likeness, membrane permeability, synthesizability, and selectivity.
With a single set of weights, LiteMol-1 can conditionally generate small molecules, peptides, cyclic peptides and more.
Related event: LiteFold Unveils LiteMol-1, a Low-Cost Multi-Molecule Foundation Model(2 posts)→
More from Research
- Fortnow: why formalizing complexity results in Lean is hard — proving polynomial time — fortnow · 2026-10-16
- Google's Willow chip cuts quantum errors 40x, raising Q-Day concerns for Bitcoin — jamestagg · 2026-10-16
- A Cookbook for Machine Learning: compiling ML goals into tractable optimization — goyal__pramod · 2026-10-16
- Transluce vets: a year of neuron autointerp work still can't explain complex behaviors — aryaman2020 · 2026-10-16
- Transformer trained on flying bat recordings yields hippocampus 'digital twin' revealing hidden structure — patrickmineault · 2026-10-16
- One year of Recursive Language Models: GPT-5-mini beats GPT-5 by 110% on 132k-token tasks, cheaper too — a1zhang · 2026-10-16