NVIDIA and University of Washington Launch RFD4-Proteina for Faster AI-Driven Protein Design
In short: NVIDIA's BioNeMo team, working with the University of Washington's Institute for Protein Design, has introduced RFD4-Proteina, a model that generates both a protein's sequence and its full atomic structure in a single pass instead of handing off to a separate inverse-folding tool. By working with a residue-level representation and a streamlined diffusion process, the model needs only about 60 denoising steps instead of roughly 400, which translates into markedly quicker runs and a much smaller memory footprint than its predecessor RFD3 when generating a 200-residue protein. During generation, the model can also run a reward-guided search that scores and branches partial designs using GPU-accelerated evaluation tools such as OpenFold3 and TMol, steering computation toward the most promising candidates. It ships both as an open-source model on GitHub and as a NIM microservice reachable over HTTP, with toolkit skills available to let automated agents chain generation, scoring and evaluation steps together.
This summary was generated automatically by AI from NVIDIA Developer's publication. It is our own text, not a copy of the original — facts, figures and quotes belong to the source, linked above and below.
What changed?
- 1Generates protein sequence and full atomic structure in one unified step, eliminating the separate inverse-folding handoff used by prior pipelines
- 2Switches to a residue-level latent representation and a re-engineered diffusion process, dropping the number of sampling steps from about 400 to 60
- 3At a 200-residue protein, this yields roughly 3.8 times quicker generation and about 84% less peak memory use compared with the earlier RFD3 model
- 4Built-in reward-guided search evaluates and refines partial designs mid-generation using CUDA-accelerated scoring tools like OpenFold3 and TMol
- 5A new BioIR execution backend adds a further 1.37x to 1.73x speedup over the standard compiled backend across tested Hopper-generation GPUs
- 6Offered as an open model on GitHub for full code control, and as a NIM microservice with an HTTP API; BioNeMo Agent Toolkit skills can automate full design-search-evaluation cycles
| Parameter | Before | Now |
|---|---|---|
| Sampling steps | ~400 (Proteina-Complexa) | 60 |
| Generation speed vs RFD3 | 1x | 3.8x faster at 200 residues |
| Peak memory vs RFD3 | Baseline | 84% lower at 200 residues |
| Sequence/structure design | Separate inverse-folding step (e.g. ProteinMPNN) required | Sequence and all-atom structure co-designed in one model |
| Backend speedup (BioIR vs default) | 1x | 1.37x-1.73x across tested Hopper GPUs |
| Deployment options | Not specified | Open model on GitHub or NIM microservice with HTTP interface |
Why it matters
For computational biology and drug discovery teams, faster, cheaper protein generation with built-in quality scoring means more design candidates can be explored within the same compute budget, shortening iteration cycles between computational design and lab testing.
Sources
- NVIDIA DeveloperOfficialPrimary sourceOriginal article →„Design Proteins with RFD4-Proteina Using Reward-Guided Inference-Time Search“21 Oct 2026, 15:30
- Published by source
- 21 Oct 2026, 15:30
- Found by our system
- 21 Oct 2026, 15:47
- Summary generated
- 21 Oct 2026, 15:48
This article was written by AI from the original source. Facts, numbers and prices come from the source; missing values are marked “Not specified”. Legal notice, copyright and privacy