Isomorphic Labs

Gated referral access to AlphaFold 3 for protein and complex structure prediction, plus a non-commercial server for academic research use.

Website
1 model · 1 with inference

Overview

Isomorphic Labs is the commercial steward of AlphaFold 3, the structure-prediction model for proteins and biomolecular complexes. It is best understood as a gated access route rather than an open runtime: there are no downloadable weights and no fine-tuning, and commercial use is arranged directly with Isomorphic. For anyone searching where AlphaFold 3 inference is officially available, this entry marks the controlled path — academic use through a hosted server, commercial use by referral.

What you can run through Isomorphic Labs

The model in scope is AlphaFold 3, which predicts the joint structure of proteins together with nucleic acids, small-molecule ligands, ions, and modified residues. It extends the AlphaFold line from single-protein folding to full biomolecular complexes, targeting interaction and assembly modeling. Access is to predictions from AlphaFold 3 itself; the model is not distributed as weights, and no custom training or fine-tuning is offered through this route.

Accessing AlphaFold 3

Two channels exist, both gated. Non-commercial and academic users run AlphaFold 3 through the hosted AlphaFold Server, which provides structure prediction for research use under its own usage terms. Commercial access does not run through a public API or self-hosted deployment; it is granted by arrangement with Isomorphic Labs, making this a referral relationship rather than a general-purpose inference platform. Teams that need open weights, on-prem execution, or fine-tuning should treat this as a pointer to the official source and look to AlphaFold3-class alternatives for a runnable model, while those who simply need authoritative predictions can use the academic server within its non-commercial scope.

Run inference on Isomorphic Labs (1)

AlphaFold 3

Google DeepMind

Released May 8, 2024

12.6K8.4K

Diffusion-based structure prediction model for biomolecular complexes, spanning proteins with DNA, RNA, small molecules, ions, and modified residues.

Protein