How to Download and Install R from r-project.org

Updated

Go to r-project.org and click Download (or the CRAN link), pick a CRAN mirror, then choose the build for your platform — Windows, macOS, or a UNIX source tarball. R is described on the site as a free software environment for statistical computing and graphics that compiles and runs on a wide variety of UNIX platforms, Windows, and macOS. The site does not publish system requirements or a step-by-step installer walkthrough, so the platform-specific steps below are general practice; the authoritative source for installation and license questions is the project's own FAQ, which the homepage points you to before emailing.

Step 1: Reach the download page

From the homepage navigation, the relevant entries are:

  • Download — the primary entry point for getting R.
  • CRAN — the Comprehensive R Archive Network; the homepage states that to download R you should "choose your preferred CRAN mirror."

Either path leads to the mirror selection. A mirror is a regional copy of the R files; the site does not recommend one over another, so pick a location near you for faster transfers.

Step 2: Choose your platform

After selecting a mirror you choose the build matching your operating system. The homepage names three platform families:

Platform What you are looking for Notes
Windows The Windows subdirectory of the mirror Download the current release installer for your architecture
macOS The macOS subdirectory of the mirror Match the build to your Mac type (Apple silicon vs. Intel); the site itself does not spell out this distinction
UNIX Source code The homepage says R compiles on a wide variety of UNIX platforms, so expect to build from source rather than run a ready-made installer

The site does not list hardware or OS version requirements, so check the README or release notes inside the platform directory on the mirror before downloading.

Step 3: Install

The r-project.org pages do not document installer screens or options, so treat the following as the normal pattern rather than site-quoted instructions:

  1. Windows — run the downloaded .exe. The wizard asks for a language, destination folder, and components; accepting defaults gives you R, the documentation, and the standard GUI.
  2. macOS — open the .pkg and follow the installer. If your Mac warns about an unidentified developer, use the standard system override for signed packages rather than disabling security globally.
  3. UNIX — unpack the source tarball, then run ./configure, make, and make install. Expect to need a Fortran and C compiler; the site does not enumerate dependencies, so read the INSTALL file shipped with the source.

Verify the result: open a terminal or the R console and run R --version (or check the startup banner). You should see the release number you downloaded.

Which version will you get?

The homepage news section is the place to confirm the current release before you download:

  • R 4.6.1 (Happy Hop) was released on 2026-06-24.
  • R 4.5.3 (Reassured Reassurer) was a wrap-up release on 2026-03-11.
  • R 4.6.2 (Diversified Portfolio) prereleases appear starting 2026-10-19, with the final release scheduled for 2026-10-29.

If you need reproducibility for a project, note the version you install; if you want the newest fixes, wait for or track the 4.6.2 schedule above.

Where to go when something goes wrong

The homepage directs installation, download, and license-term questions to the project's FAQs before contacting anyone by email. The Documentation section links manuals, and Getting Help covers support channels. For packages beyond base R, the site lists Bioconductor, R-Forge, and R-Hub under Other Links; these are separate from the base download and are not required to install R itself.

One practical note: the site presents R as free software, but it does not state pricing or login terms for any download path, so do not assume a mirror offers accounts or paid tiers — the mirror is simply a file host.