Apple's Safari-vs-Chrome Claims Rest on Prerelease Builds
Software / news
Apple's Safari-vs-Chrome Claims Rest on Prerelease Builds
The footnotes on Apple's new Switch to Safari page show tests run in August 2026 on one laptop, against two different Chrome builds.

Apple's new Switch to Safari page says Safari is up to 35 percent faster than Chrome at loading frequently visited sites and streams video for up to five hours longer on a Mac. Both figures come from Apple's own testing in August 2026 on prerelease software, according to the footnotes on the page.
9to5Mac reported the campaign on Oct. 12. It also includes two YouTube videos, on Distraction Control and Intelligent Tracking Prevention. Apple describes Safari there as "a more powerful and private browser than Chrome."

What the footnotes say about the 35 percent and the five hours
Footnote 1, the battery claim, used a 15-inch MacBook Air with an M5 chip and 16GB of memory. The Mac ran prerelease macOS "Golden Gate" and prerelease Safari 27.0. It played 1080p video on Wi-Fi with the display at eight clicks from the bottom.
The Chrome in that test was v151.0.7922.76. Apple's own footnote says battery life varies.
Footnote 3, the page-load claim, timed snapshots of 12 popular websites under simulated network conditions on the same laptop. The Chrome there was v151.0.7922.138, a different build from the one in the battery test. The page does not name the 12 sites.
- Safari18.67 % of battery
- Chrome17.33 % of battery
Source: Michael Tsai's blog quoting Matt Birchler, accessed 2026-10-14
Where the page and an outside test disagree
Blogger Matt Birchler ran the same kind of comparison in 2024 on a MacBook Pro after a clean reinstall of macOS Sonoma. Michael Tsai's blog quotes the result: Safari averaged 18.67 percent drain per three-hour test and Chrome 17.33 percent, about 9 percent less for Chrome. Chrome won five of six direct comparisons.
Birchler wrote that run-to-run variability was a caveat. In a later update he said controlled tests showed very close results, with Chrome using slightly less battery, and that nobody had sent him 2024 data showing Chrome was much worse.
That is one person, one Mac and an older Safari, so it does not refute Apple's M5 figure. A shader-hashing fix in Unreal Editor shows the same caution applies to any speed figure: its 12.90-to-1.59-second gain was summed CPU time across threads, not wall-clock startup. It does show the five-hour claim is an "up to" number from a single configuration, not a general property of the browser.
A benchmark footnote that lists hardware Safari cannot run on
Apple's headline calls Safari "the world's fastest browser", citing JetStream 3.0, MotionMark 1.3.2 and Speedometer 3.1. Footnote 4 says the values are rounded and that Safari scored higher on average against Chrome v151.0.7922.138, Edge v151.0.4129.78 and Firefox v153.0.4.
The hardware list for those tests includes an Intel Core Ultra X7 PC running Windows 11 Home. Footnote 6 adds Snapdragon 8 Elite Android phones. The same page says Safari runs only on macOS, iOS and iPadOS.
The page does not say which browsers ran on which machine, and it publishes no scores. Google's Gemini 4 Argon launch, covered here, had the same shape: vendor-supplied figures that no third party had reproduced at launch. The 12 sites, the raw timings and the per-benchmark results are all absent.
| Claim on the page | Basis in the footnote | Left out |
|---|---|---|
| Up to 5 hours more streaming | One M5 MacBook Air, Chrome v151.0.7922.76 | Number of runs |
| Up to 35% faster page loads | 12 sites, simulated network | Site list, timings |
| World's fastest browser | JetStream, MotionMark, Speedometer | Scores, per-device results |
What the comparison chart leaves out
The page's comparison table marks Safari "Yes" and Chrome "No" on all nine rows, including "Lets you hide annoying distractions and remembers your preferences" and "4K HDR video playback on Mac for YouTube, Netflix, and Apple TV." No row gives Chrome a pass.
Apple's Notify Me and Organize by topic features run on Apple Intelligence, which needs a recent device and one of 16 supported languages. Apple's footnote adds that some server-side features have daily usage limits.
Safari 27 and macOS Golden Gate were prerelease when Apple ran its tests, and 9to5Mac's report carries no response from Google.
For a reader weighing a switch, the check that matters is cheap: run the same 20 minutes of work in both browsers on battery and compare the percentage left. That is the test Birchler published, and the page offers nothing to stop anyone repeating it.
Sources
More in Software
- 01Rivet, a YC W23 Company, Runs Agent Actors on S3 and Claims 72 KB EachThe Apache 2.0 orchestrator says an actor cold-starts in 4.99 ms at p99. Every figure is Rivet's own, and the Launch HN thread questioned both the premise and the AWS bill.
- 02GitHub Actions Lockfiles: What gh actions-lock Pins and What It Leaves LooseWorkflow lockfiles went generally available on October 28. The extension's own README lists four gaps, including Docker images and reusable workflows in other repositories.
- 03C++26 Reportedly Fails Its ISO Ballot, With Contracts the Sticking PointSecondary reports put the October 23 vote at 12 yes of 20 and nine no votes of 22 bodies. The numbers do not reconcile, and no ISO or WG21 record has been published to settle them.
- 04A Postgres Benchmark Found jsonb Read One Key 2.7 to 10.6 Times Faster, and Larger on Disk in 6 of 11 CasesChris van Eijk of Now-Next loaded 1 million small documents and 5,000 large ones into PostgreSQL 17.11 and 18.6. The speed result is consistent. The storage result depends on whitespace and the 2 kB compression threshold.