(last updated on )
Written by on CSS Wizardry.
Independent writing is brought to you via my wonderful Supporters.
For the casual performance enthusiast, dedicated device testing is likely to be overkill. However, anyone working full time on site-speed will probably benefit from having at least one real device to hand.
I’m working on a project at the moment where mobile INP is our key focus, and while Chrome Desktop’s DevTools has been a great starting point, replicating real-world mobile interactions and performance has been a consistent challenge.
I decided to deep dive and look at what would constitute a real low- and mid-tier mobile device in 2025, and how that maps to the Chrome DevTools presets.
When assessing, I wanted to factor in a sensible mix of:
And remember, as we’ll look at next, trying to triangulate on one single device per category is a little futile in itself—sheer device diversity and evolution means that anything we pick is going to be something of a compromise. My aim, however, is to minimise that compromise as much as possible and distill our choices down to a single representative device for each category. Can I do it?
For the longest time, the Chrome team touted the Motorola Moto G4 as the global baseline device. As a result, I bought a physical Moto G4 and used that for my real-device testing, inspecting and tracing Chrome remotely. However, seeing as the G4 was released over nine years ago and is currently locked to Android 7.0, that soon stopped being entirely representative.
If you’d like to learn more about testing with real devices, either Android or iOS, then arrange a Masterclass.
In Lighthouse 10, the team switched to an emulated Moto G Power (though this was more of a viewport and DPR change than hardware capabilities).
The key word is here emulated: there is no real device and all of the network and CPU throttling is simulated, so actually listing any device at all, to me at least, seems a little disingenuous. There is no Moto anything, so they’d be better off saying ‘low-end mobile’, or words to that effect.
In Chrome’s DevTools, that’s exactly what they do! Low-tier mobile defaults to a 6× CPU slowdown and a 3G-like connection. Mid-tier mobile is a 4× CPU slowdown and a Slow 4G-like network. If you take the time to calibrate these presets yourself, you’ll get something a little more bespoke. For me, on my development machine, a mid-tier CPU is 2.9× slower than my machine’s true capabilities and low-tier mobile is a staggering 9.1× slower. That’s quite a departure from the off-the-shelf 4× and 6×!
It’s important to remember that ‘low-tier’ does not mean old. Likewise, ‘mid-tier’ does not mean ‘a flagship from 2018’. They are a device class in their own right and it’s perfectly possible for your users to be using brand new low-tier devices. As such, we needn’t look to the past for our benchmarks.
With that in mind, let’s go find ourselves some real low- and mid-tier devices that we can use for real-world testing in 2025.
TL;DR: If you want broadly representative Android hardware for real-world web performance testing in 2025, buy a Samsung Galaxy A15 5G (SM-A156x) (low-tier) and a Samsung Galaxy A54 5G (SM-A546x) (mid-tier). Both are mass-market, widely ranged across regions, have multi-year software support, and map well to Chrome DevTools’ Low- and Mid-tier presets.
Based on these factors alone, many devices might make the grade, so let’s look in a little more detail.
The A15 is one of the cheapest mainstream 5G handsets Samsung markets worldwide: you’ll find it in Latin America, Africa, Europe, South Asia, and North America. That ubiquity is quite rare; most lower-end Androids tend to be regional brands (Infinix, Tecno, Lava, etc.).
The A15 also reflects what a budget-conscious, mainstream consumer gets in 2024/25: limited CPU/GPU headroom, middling storage I/O, and memory constraints. Perfect for showing how web performance characteristics change when you head toward the floor of mass-market devices.
While not even ‘flagship-lite’—its Exynos 1380 chip and GPU are properly mid-market—the A54 is one of the world’s top-selling mid-tier Androids, and it has staying power in both carrier contracts and SIM-free contexts.
Further, Samsung sells essentially the same A54 device everywhere, which makes test data reproducible between Europe, India, Southeast Asia, Latin America, and even the US. Many competitors (e.g. Xiaomi, OPPO, realme) tend to fragment their mid-tiers by market.
Finally, both the A15, and A54 get Samsung’s long software support, meaning they will likely still be in market for several years.
I actually went out and bought myself an A54 5G after conducting this research!
If you’d like to replicate these devices in WebPageTest, you’ll need to pick from their (slightly dated) device emulation list. While there aren’t any one-to-one matches for current hardware, there are some pretty close proxies. I’d recommend:
It’s worth stressing that these aren’t perfect matches (they never will be), but they’re close enough to give you realistic lab results when you can’t test directly on a physical device.
The web is used on billions of people, but not all devices are created equal. If we’re serious about building a fast, accessible web, we need to calibrate our benchmarks against hardware that people actually own and not just the flagships we keep in our pockets and the presets that DevTools hands over.
Having just one real device in your arsenal is a great place to start, and in 2025, that’s the Samsung Galaxy A15 5G for low-tier, and the Galaxy A54 5G for mid-tier testing.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。