AI Bio Atlas LogoaibioatlasSEARCH • COMPARE • CHOOSE
Home›AI Drug Discovery›MoLFormer
drug-discoveryDrug Discovery & Chemistry
AI Drug Discovery · Molecular Foundation Models

MoLFormer

By IBM Research →

Large-scale chemical language foundation model for molecular property prediction

MoLFormer is a chemical foundation model trained on 1.1 billion molecules with linear attention transformers, achieving state-of-the-art accuracy in quantum chemical properties and ADMET prediction.

Trained on 1.1 billion SMILES strings from PubChem and ZINC databasesLinear attention mechanism enabling efficient scaling to large molecular graphsState-of-the-art benchmarks on MoleculeNet across physical chemistry and biophysics tasksFine-tuning support for custom pharmaceutical assay endpoints
Compare

ℹ️ No direct buying or access link is currently recorded. This product remains available for research and comparison.

Report incorrect information or a broken link →

Overview

MoLFormer is a chemical foundation model trained on 1.1 billion molecules with linear attention transformers, achieving state-of-the-art accuracy in quantum chemical properties and ADMET prediction.

Information checked against an official source; not a hands-on test. Source · Last reviewed: 20/09/2026, 11:09:40

Key Features

  • Trained on 1.1 billion SMILES strings from PubChem and ZINC databases
  • Linear attention mechanism enabling efficient scaling to large molecular graphs
  • State-of-the-art benchmarks on MoleculeNet across physical chemistry and biophysics tasks
  • Fine-tuning support for custom pharmaceutical assay endpoints
Interactive 3D Structure

MoLFormer Predicted Complex

Streams real 3D atomic coordinates from RCSB Protein Data Bank
⇄ Drag to rotate · Scroll to zoom
pLDDT / B-Factor:
>90 Very high 70-90 Confident 50-70 Low <50 Very low

Academic Context & Research Evidence

Biological & Workflow Fit

Biological Application
Solubility, toxicity (hERG, Ames), BBB permeability, and binding affinity prediction
Research Workflow
SMILES input -> high-dimensional latent molecular embeddings -> downstream property predictions
Compute & Hardware
1x GPU or CPU for inference; multi-GPU for fine-tuning
Licensing & Academic Use
MIT License
Documented Evidence
View validation publication / source ↗

Cite this Tool

Use this citation format when referencing MoLFormer in scientific publications and benchmark papers.

@software{molformer_2026,
  title = {{MoLFormer}},
  author = {{IBM Research}},
  year = {2026},
  url = {https://github.com/IBM/molformer},
  note = {Indexed on aibioatlas - AI for Biology and Drug Discovery}
}

Peer-Reviewed Literature & Preprints

Live scientific citations streamed from Europe PMC and PubMed for MoLFormer.

⏳ Fetching real-time literature from Europe PMC & PubMed...

Technical / Product Information

Missing values mean the catalog has no recorded information. They do not mean a feature is absent.

Entry typeAI Model
Access modeOpen Source
AI roleChemical Foundation Model
Input dataNot recorded
Output dataNot recorded
Licence conditionsMIT License
Commercial eligibilityOpen source release
Compute requirements1x GPU or CPU for inference; multi-GPU for fine-tuning
ValidationNot recorded
TypeChemical transformer foundation model
Intended useNot recorded
CompatibilityNot recorded
ManufacturerIBM Research
Biological applicationSolubility, toxicity (hERG, Ames), BBB permeability, and binding affinity prediction
Research workflowSMILES input -> high-dimensional latent molecular embeddings -> downstream property predictions
Evidence levelPeer-reviewed research (Nature Machine Intelligence 2022)
Integration evidencehttps://github.com/IBM/molformer
Laboratory handoffFilters high-risk toxic molecules before compound synthesis
AvailabilityAvailable on Hugging Face and GitHub
Price / accessFree Open Source

Research fit & compatibility

No software–hardware integration has been verified for this entry yet. Explore documented research workflows.

Gallery & resources

No licensed product photograph recorded for MoLFormer yet. Visit the official resource below for manufacturer imagery.

BEFORE YOU CHOOSE

Evaluate AI Drug Discovery

  • Which input, output and scientific task does the selected version support?
  • Is this a model, supporting tool, commercial platform or research prototype?
  • Verify the code, weights, data licence and independent validation before selecting a workflow.
Selection guide →Find alternatives →

Alternatives

Similar entries from the same category.

drug-discoveryDrug Discovery & Chemistry
End-to-End AI Drug Discovery Platform

Pharma.AI

Commercial Platform · Commercial Subscription & Co-Development

Integrated generative AI platform from target discovery to clinical candidate design

Insilico Medicine

Enterprise partnership — contact vendor
Explore →＋ Compare
drug-discoveryAI Drug Discovery
Retrosynthesis

ASKCOS

Research software / framework · Public code / project terms

Computer-aided organic synthesis planning

MIT / MLPDS Consortium

See project terms; compute costs may apply
Explore →＋ Compare

FAQ

Where is this product available?

Available on Hugging Face and GitHub

How is pricing handled?

Prices reflect the source at its last check. Confirm current pricing and regional availability on the official site.

Share this entry with your lab or research teamSpread reproducible science across your network
𝕏 Postin LinkedIn

Peer Reviews & Community Ratings

Feedback from researchers and computational biologists evaluating MoLFormer.

5.0
★★★★★Based on 0 researcher evaluations
Biological Accuracy
4.8/5
Ease of Installation
4.3/5
Documentation & Code
4.6/5