RFD4-Proteina
Flow-matching model from the Baker lab and NVIDIA that co-designs protein sequence and all-atom structure, covering binders, antibodies, enzymes and small-molecule pockets, with reward-guided search.
| Category | Protein structure & design |
|---|---|
| GitHub | RosettaCommons/RFD4-Proteina |
| First public | 2026-10-19 (source) |
| Paper | https://rfd4-proteina.org/assets/paper/RFD4-Proteina.pdf · Preprint (2026) |
| Code license | BSD-3-Clause |
| Weights license | Hugging Face weights BSD-3-Clause |
| What this allows | Permissive (commercial OK) Use, modify and ship it inside closed-source commercial products; keep the license text and copyright notices. What do these mean? |
| License source | huggingface.co/nvidia/RFD4-Proteina · checked 2026-10-22 |
| Pricing | Open source |
| Tags | #protein-design#flow-matching#all-atom#antibody |
| Last checked | · links verified |
Can I use RFD4-Proteina commercially?
Yes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices.
- Code: BSD-3-Clause
- Weights: Hugging Face weights BSD-3-Clause
- Source: huggingface.co/nvidia/RFD4-Proteina · license checked 2026-10-22
What the license types mean · not legal advice.
Specs
| Task | Monomers, protein/DNA binders, antibodies and nanobodies, enzymes, motif scaffolding, small-molecule binders, partial diffusion, inverse folding |
|---|---|
| Modalities | Proteins, nucleic acids, small molecules (unified autoencoder) |
| MSA required | — |
| Open weights | Yes (Hugging Face nvidia/RFD4-Proteina) |
| GPU need | NVIDIA GPU, CUDA 12/13, Linux |
| GitHub stars | 333 |
| Interface | Package (pip install rfd4-proteina), CLI, Python API, Hugging Face Diffusers, NVIDIA NIM |
Compare with…
Pick 1–3 other Protein structure & design tools to compare side by side with RFD4-Proteina.
Related
DeepMind's open-source protein structure prediction model (monomer and multimer).
Structure prediction of proteins, nucleic acids, ligands and their complexes; hosted web server plus open code, weights on request for non-commercial use.
Open biomolecular interaction models (Boltz-1, Boltz-2) for complex structure and binding-affinity prediction.