Google DeepMind has made the code and weights for AlphaFold 3 available to researchers
After facing criticism for undermining the reproducibility and transparency of science, Google DeepMind has released the source code for AlphaFold 3 for anyone to access. Additionally, researchers with an academic affiliation can request access to the model's training parameters.
After initially releasing AlphaFold 3 through a web server that limited the types of predictions that could be performed with the tool, Google DeepMind has uploaded the AlphaFold source code to a GitHub repository for anyone to download. Researchers with an academic affiliation can request access to the model parameters, although it is up to Google DeepMind to decide whether it approves the requests. Like the web server tool, the AlphaFold 3 code and parameters are only licensed for some non-commercial applications.
Google DeepMind justified its decision to make AlphaFold 3 through a web server by describing the strategy as one that encouraged and supported access to research while also protecting DeepMind's commercial interests. Isomorphic Labs, a DeepMind spin-off, is leveraging AlphaFold 3 as part of its drug discovery pipeline. The move was quickly criticized as going against the transparency and reproducibility that characterize scientific research.