TL;DR
Get everyday helpers delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Zerobrew, a package manager for macOS and Linux, published benchmarks claiming substantially faster installs than Homebrew in a test of 100 packages. The overall result was 6.6x faster for cold installs and 68.3x for warm installs; 24 packages were at least 100x faster in the warm test. The results are developer-run benchmarks, and Zerobrew does not yet run Homebrew post-install steps.
Zerobrew, a package manager for macOS and Linux, has published tests reporting faster installations than Homebrew across a 100-package run, including a 68.3x speedup for warm installs. The project’s “100x faster” tagline refers specifically to 24 of the 100 packages in its warm-install test, not to the full package set or every installation.
The benchmark reports that installing all 100 packages took 775.60 seconds with Homebrew and 117.43 seconds with Zerobrew in the cold test, a reported 6.6x speedup. In the warm test, Homebrew took 637.50 seconds and Zerobrew took 9.33 seconds, a reported 68.31x difference. Zerobrew defines a cold run as having no packages installed and an empty download cache. For a warm run, the packages are still uninstalled, but downloads from the cold run remain cached.
The results were measured on October 8, 2026, using Zerobrew 0.3.5, described as a development release, and Homebrew 7.0.8. The test machine was a MacBook Pro with an M3 Pro processor and 18 GB of memory, running macOS 26.6.2, with an approximately 318 Mbit/s internet connection. Zerobrew reports a median speedup of 5.3x in cold installs and 56x in warm installs. The fastest cited warm result was ca-certificates at 252.12x; the slowest cited cold result was go at 1.61x.
The project says warm installs are faster partly because its store retains both the downloaded package archive, or bottle, and the unpacked package after an uninstall. Reinstalling can then clone the stored files rather than unpacking them again. That also uses more disk space. Zerobrew’s benchmark notes another difference: it did not run post-install steps, while Homebrew did. The project lists 13 formulae for which Homebrew performed such steps during these tests, including ca-certificates, node and Python.
What the Speed Gap Means
The results point to a potential time savings for developers who repeatedly install and remove command-line tools, libraries and language runtimes. The difference is most pronounced when package files are already cached: Zerobrew says its retained unpacked packages let a warm reinstall finish without the same extraction work. For a user setting up many packages, the reported 9.33 seconds versus 637.50 seconds in the 100-package warm test is a material difference if the test conditions match their use.
But the headline figure needs its test conditions attached. The 100x figure is not the overall benchmark: it applies to 24 individual packages in the warm run. Cold installs were much less dramatic, with the overall test showing 6.6x, and large downloads can limit the difference when both tools are waiting on the network. The comparison also does not establish that Zerobrew completes precisely the same work, since the project says it has not implemented the post-install steps Homebrew ran for some formulae.
For people choosing a package manager, speed is only one consideration. The published results do not answer questions about compatibility across a broad range of systems, long-term reliability, package coverage, security practices or the effects of using additional disk space. Readers should treat the numbers as a useful, project-published performance snapshot, rather than an independent assessment of those broader trade-offs.
As an affiliate, we earn on qualifying purchases.
How Zerobrew Differs From Homebrew
Zerobrew describes itself as bringing a “uv-style architecture” to Homebrew packages on macOS and Linux. Its command-line interface supports installing and uninstalling packages, upgrading selected or outdated packages, installing from a Brewfile and running a package without linking it through a command beginning with zbx. The project’s README provides both a standalone installer and installation through a Homebrew tap.
The benchmark page makes clear that installation speed depends on the test setup. The project also reports a separate run on an approximately 69 Mbit/s connection, where the overall comparison was 3.3x faster cold and 69x faster warm. It says large packages such as go and llvm spend most of their cold-install time downloading, leaving less room for the package manager’s handling of local files to change the result. These figures are the project’s own measurements; the supplied material does not include an independent replication.
The benchmark’s stated scope is narrower than a general claim about all packages and machines: it records one set of results on a specified Mac and operating-system version, with particular software releases and network conditions. The result may differ with other hardware, connection speeds, package selections or cache states.
“zerobrew brings uv-style architecture to Homebrew packages on macOS and Linux.”
— Zerobrew project documentation
fast command line package installer
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What the Benchmark Leaves Open
The published figures do not establish how Zerobrew performs across different Macs, Linux systems, internet connections or package collections. The project labels version 0.3.5 a development release, and the material provided does not include results from an independent benchmark or a direct test of reliability over time. It is also unclear how much extra storage the retained bottles and unpacked packages use in typical day-to-day use.
Functional equivalence remains a limitation in the comparison: Zerobrew says it does not yet run post-install steps, while Homebrew did so for 13 formulae in the benchmark. The project names rebuilding the certificate bundle as one example. The available results do not show a follow-up run in which both tools perform the same post-install work, or quantify how the timings would change if that work were included on the Zerobrew side.
Nor does the 100x label describe every package. Zerobrew says 24 of 100 packages reached that threshold in the warm test; its reported warm speedup ranged from 17.92x for go to 252.12x for ca-certificates. The underlying results are a snapshot under the specified test conditions, and the comparison basis for any use outside those conditions remains unknown.
As an affiliate, we earn on qualifying purchases.
The Next Benchmark Milestone
The next useful comparison would test Zerobrew and Homebrew with equivalent post-install steps, alongside repeated runs on more than one machine and network connection. Zerobrew’s documentation says it plans to run Homebrew with –skip-post-install in its next comparison until Zerobrew can run those steps too. That would make the package-manager timing comparison more like-for-like, although it would not itself measure the time needed to complete the skipped work.
Readers can also watch for updates to Zerobrew’s release and benchmark results as the project develops. The current material does not specify a date for the next results or set out a plan for independent testing. Until additional evidence is published, the reported figures should be read as measurements of the stated setup—not a guarantee of the same speedup on another system.
software installation speedup tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Is Zerobrew 100 times faster than Homebrew overall?
No. The project reports 6.6x faster cold installs and 68.31x faster warm installs for the full 100-package test. Its 100x tagline refers to 24 individual packages that reached at least that speedup in the warm test.
What does Zerobrew mean by a warm install?
In its test, the packages were uninstalled but downloads from the earlier cold run remained cached. Zerobrew says it also keeps unpacked package files in its store, allowing a reinstall to clone them. The project notes that this approach uses more disk space.
Were the speed tests independently verified?
The supplied material presents results published by Zerobrew. It does not provide an independent replication, so the results should be understood as the project’s own benchmark measurements.
Did both package managers perform the same work?
Not fully, according to Zerobrew. The project says it did not run post-install steps, while Homebrew did for 13 formulae in the test. That limits how directly the reported timings compare total work completed.
Which systems does Zerobrew support?
The project describes Zerobrew as supporting macOS and Linux. The published benchmark detailed in the supplied material was run on a MacBook Pro with an M3 Pro chip, not across a range of systems.
Source: hn
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
