Updated May 2026

Mobile Internet Speed: 4G vs 5G Explained

Mobile speed tests are the most unpredictable of any connection type. The same phone on the same carrier can test at 5 Mbps in one spot and 300 Mbps a few blocks away. Your phone can show a 5G icon while running slower than LTE would in the same location. If you want to make sense of your mobile speed test results, you need to understand what's actually driving your phone's connection performance.

4G LTE vs 5G: What the Numbers Actually Mean

Factor 4G LTE 5G Sub-6 GHz 5G mmWave
Typical download 10–100 Mbps 50–300 Mbps 500 Mbps – 4 Gbps
Typical upload 5–50 Mbps 20–100 Mbps 100–500 Mbps
Typical latency 30–70ms 10–40ms 5–20ms
Coverage Almost everywhere Cities and big suburbs Small pockets of dense cities only
Building penetration Good Good to moderate Very poor (blocked by windows)
Real-world reliability Consistent where covered Mostly consistent; varies by band Very hit or miss (range ~150m)

Understanding 5G Variants

The "5G" label on your phone can mean really different things. Carriers put 5G on different frequency bands, and each one has different trade-offs between speed and coverage.

Sub-6 GHz 5G (the most common)

This is what most people actually have. It uses frequency bands below 6 GHz — similar range to 4G LTE but with newer, smarter tech. You get better coverage, decent signal inside buildings, and a real speed boost over LTE. In most places, sub-6 GHz 5G hits 50–300 Mbps. That's noticeably faster than LTE, but it's not the gigabit speeds you see in commercials.

mmWave 5G (millimeter wave)

This is the crazy-fast 5G you see ads about. It can hit multiple gigabits per second. But there's a big catch: it barely works outside a 100–300 meter range from the antenna. Walls, glass, rain — even people standing between you and the tower block it. You'll find it in some dense city areas near specific antennas. Most people will never actually use mmWave in their daily life.

Low-band 5G (below 1 GHz)

Some carriers run 5G on the same low frequencies they used for older 4G LTE. Coverage is great and it goes through buildings well. But the speed boost over LTE is tiny — often just 10–30% faster. If you see a 5G icon in a rural area, this is probably what you're on.

The "5G E" problem

AT&T used to label regular 4G LTE Advanced as "5G E" on phones. That's pure marketing. "5G E" is not 5G at all. It's just a faster version of LTE. A 5G icon on your phone doesn't guarantee you're on real 5G technology.

What Actually Affects Your Mobile Speed Test

Signal strength (the biggest factor)

Signal strength is measured in dBm — a scale where less negative means stronger. Your phone shows "bars" instead of the actual number, but here's roughly what each bar means:

Bars / Signal level Approximate dBm Expected performance
4–5 bars (excellent) -50 to -70 dBm Full speed for your network type
3 bars (good) -70 to -85 dBm Good speed; close to the max
2 bars (fair) -85 to -100 dBm Noticeably slower; higher latency
1 bar (poor) -100 to -110 dBm Slow and unreliable; calls may drop
No service / SOS only Below -110 dBm No usable data connection

Key takeaway: strong 4G LTE almost always beats weak 5G. The technology doesn't matter as much as the signal quality when you're far from a tower or inside a thick-walled building.

Network congestion

Think of a cell tower like a water pipe shared by everyone on your block. During a concert or a crowded airport, everyone's fighting for the same pipe. Your speed can drop 90% compared to testing at 5 AM in the same spot. Mobile congestion hits even harder than cable internet congestion because a single cell sector has way less total capacity than a fiber node.

Carrier policies and deprioritization

A lot of phone plans have a "premium data" limit — a monthly cap where, after you've used that much data, your traffic can be slowed down when the network is busy. It's not like a hard cutoff. It only kicks in during congested times. In practice, your speed tests might look great at 5 AM but crawl at peak hours once you've hit your premium data limit for the month.

Your device and its modem chipset

Different phones support different bands and different modem generations. An older flagship from 2019 might technically be "5G" but only on a few basic bands. A 2024 flagship supports all 5G bands including mmWave. Your phone's hardware sets a ceiling on what speeds are possible — the network can't give you more than your device can handle.

How to Run an Accurate Mobile Speed Test

  1. Turn off Wi-Fi first. Go to Settings → Wi-Fi and turn it off. Wi-Fi and mobile data are two different networks. Make sure your test is actually going through cellular.
  2. Test outside. Buildings absorb cellular signals. Testing outdoors gives you a clearer picture of what your carrier's network can actually do.
  3. Test near a window if you can't go outside. Avoid basements and rooms in the middle of big buildings.
  4. Note your signal strength (bars, or dBm if you know how to pull it up on your phone) alongside the result.
  5. Test at different times. Run the test in the morning and in the evening to see how much congestion hurts you.
  6. Test in the places that matter — your home, your office, your commute route. Mobile performance is very location-specific.
  7. Compare carriers by trying a temporary SIM card from another carrier at the same location.

Mobile Speed and What You Can Actually Do With It

Activity Minimum mobile speed needed Latency requirement
Web browsing 1 Mbps Under 200ms
SD video streaming 3 Mbps Not critical
HD video streaming (1080p) 5–8 Mbps Not critical
4K video streaming 15–25 Mbps Not critical
Video calls (Zoom, Teams) 5 Mbps symmetric Under 150ms; low jitter is critical
Mobile gaming 5–10 Mbps Under 80ms preferred
Mobile as home internet hotspot 25+ Mbps Depends on household use
Mobile speed tests are a snapshot of one moment in one location. Results change a lot depending on whether you're indoors or outdoors, time of day, how far you are from the tower, and how busy the network is. Run a few tests to get a real picture of what you typically get.