Indeed, Chrome was about 20% faster running in Incognito mode without any extensions.
It's actually faster than Safari on the same 16" MBP with M1 Pro running Sonoma 14.4.
Chrome:
Safari:
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple, Google, and Others Announce Speedometer 3.0 Browser Benchmark
- Thread starter MacRumors
- Start date
- Sort by reaction score
It's actually faster than Safari on the same 16" MBP with M1 Pro running Sonoma 14.4.
Chrome:
Safari:
Wouldn’t that be expected?
A MacBook user might get higher numbers plugged into the wall in a cool room compared to on battery outside in the summer sun.
An iPad user could similarly see different results depending on the device heat status and etc etc
Lots of external factors that influence the real time performance envelope for our devices. Then you have the silicon lottery giving slight variations in chips between devices.
With a speedometer going all the way up to 120, 130 and even all the way up to 140,
….All these scores demonstrate that there is plenty of work to do, and real possibilities, to make ANY browser really Truly Snappy !!
….All these scores demonstrate that there is plenty of work to do, and real possibilities, to make ANY browser really Truly Snappy !!
Last edited:
Pretty big deal, this should lead to performance enhancements in the future.
I think the only reason Apple endorsed this gimmick tool is to convince everyone that they need to upgrade their machines!
Without making any particular effort to optimize:
M1 Max MPB = 27.0 on Safari, 27.3 on Chrome
iPhone 15 Pro = 27.5
Apple Watch S9 = 2.71 (using μBrowser)
The AW CPU is if I'm not mistaken based on the efficiency cores of the A15, and I'm really curious how it compares to some older devices.
M1 Max MPB = 27.0 on Safari, 27.3 on Chrome
iPhone 15 Pro = 27.5
Apple Watch S9 = 2.71 (using μBrowser)
The AW CPU is if I'm not mistaken based on the efficiency cores of the A15, and I'm really curious how it compares to some older devices.
No, you (and most of the users on the forum, to be fair) are misunderstanding what this is. It’s not a browser speed test, it’s a website performance test. It’s for web devs to test the performance of their website in multiple browsers.
Edit: Looks like I was partially wrong. It IS about browser performance, but it’s also about how different workflows used by web developers (and different things websites can do, and different libraries websites can use) perform on your hardware in your browser of choice, I think. It’s something more developer oriented, but the reason it’s a news story here is because it’s finally an open standard all the browser companies are contributing to (as opposed to this thing the WebKit team made that they had some input from the Blink team on). That does make it newsworthy, though perhaps it would be better on a site like ArsTechnica or even TheRegister (ie something more developer oriented) than a more consumer-y site like MacRumors.
Last edited:
It's funny to see people confidently declaring flaws when it's obvious they didn't even read the second paragraph of the article, which makes it abundantly clear that this benchmark was not developed by Apple, it was developed as a collaborative effort between all three major browser engine developers.
From the press release on the Speedometer site:
If it's doing something to make WebKit look better, that's only because the Blink and Gecko devs let them. Not to mention, if Apple was trying to make themselves look "snappier" here they did a garbage job, as Safari on my Mac scores a 27.0 and Chrome 27.3; Firefox is 24.1.
Edit to add: Here's the Microsoft Edge dev team explicitly promoting Speedometer 3.0, and encouraging you to try it out, and the Chromium official blog is doing the same.
Last edited:
I object, the speedometer is clearly reading in MPH (megapages-per-hour?).
It makes a great deal of sense that even the highest-performance modern computers would be scoring relatively low on its scale, since this is presumably going to remain in use for a while so you'd want to leave plenty of overhead for future CPUs and browser improvements. Having a finite scale pegged from day one would be pretty silly.
M1 Air 16GB on battery (93% charge)
Latest versions of all browsers, had all 3 open simultaneously, and had other tabs open (but sleeping) in Edge while Safari and Firefox had only Speedometer open.
For funzies, I ran all 3 benchmarks on browserbench.org in all 3 browsers installed on my machine. Below are the results.
Higher is better in all tests
Latest versions of all browsers, had all 3 open simultaneously, and had other tabs open (but sleeping) in Edge while Safari and Firefox had only Speedometer open.
For funzies, I ran all 3 benchmarks on browserbench.org in all 3 browsers installed on my machine. Below are the results.
Higher is better in all tests
| Speedometer 3.0 | Jetstream (Javascript test) | MotionMark 1.3 (Graphics Test) | |
| Edge | 19.9 | 307.242 | 3947.27 @ 60fps |
| Firefox | 26.5 | 215.174 | 1590.16 @ 60fps |
| Safari | 24.2 | 321.523 | 5206.04 @ 60fps |
Last edited:
That's interesting, I got slightly slower results in Chrome in Incognito mode than "regular", although realistically both of those figures and Safari are all within 2% of each other and well within the declared margin of error on any of the tests.
Which is to say that small variations almost certainly aren't meaningful unless you get really consistent results across a bunch of runs (which I at least don't, and even then 2% variation seems pretty trivial on a test that says it has a margin of ~ ±5%).
More broadly, the extreme variation of results even on similar CPUs makes it fairly obvious that there are other variables that can heavily impact browser performance, although I'm not sure what they are--lack of free RAM, extensions, many tabs open, other processes running, heat, and more are all possibilities.
If anything I think the biggest takeaway to me is that if I ran a Safari test and consistently got around 12 when other people with similar hardware and Safari/OS version got around 27, that tells me something is wonky with either my install or environment, and things are running very slowly as a result.
That said, I at least didn't get anything unexpected; my numbers are in line with the high end of what others are posting for an M1 Max (~27), Safari and Chrome are basically the same, and Firefox is only modestly slower than either, all of which have a BUNCH of windows open.
iMac 5K 27" 2017
3.4 GHz Quad-Core Intel Core i5
Radeon Pro 570 4 GB
32 GB 2400 MHz DDR4
macOS Ventura 13.6.5
Chrome: 8.70
Safari: 9.63
Meh, why so low scores? Just because it's Intel?
It's also a corporate computer with JAMF installed, could that slow it down?
3.4 GHz Quad-Core Intel Core i5
Radeon Pro 570 4 GB
32 GB 2400 MHz DDR4
macOS Ventura 13.6.5
Chrome: 8.70
Safari: 9.63
Meh, why so low scores? Just because it's Intel?
It's also a corporate computer with JAMF installed, could that slow it down?
Google's thought behind that: If competing browsers come with sophisticated ad blocking, that might slow them down.
Not necessarily because it's an Intel machine, but more likely because it's a 7 year old machine. CPU core performance in Macs made a big jump with the release of the M1 and after that Intel and AMD significantly upped their games too. A base M1 Mac is about twice as fast as your machine and an M3 Max can be as much as 10 times faster in certain benchmarks.
And no matter how fast our processors are, when we see those results, we immediately think we have a slow computer.
So we want to upgrade.
And give more money to those companies.
Scores are bad for me as well with Windows 11 on Chrome and Edge with a score just over 10, and this is with an AMD 5800X3D, 4090 and 64GB 3200.
Edit: Tried a fresh install of Firefox and that scored 15.4, so that's big improvement.
JetStream 2 out of curiosity scores 282 on Chrome and 183 on Firefox, so the opposite performance this time 😕
Last edited:
After updating to from 17.4 to 17.5, I got a 1.8 point boost on my M1 Max 64GB
