Guide

VLESS on a 5G Mobile Proxy: Setup and a Fair Speed Check

Published Sep 26, 2026 · Updated Oct 17, 2026

The first question a speed-minded customer asks about an encrypted tunnel is whether it costs throughput. The honest answer is that you should measure it on your own line and your own connection, and this guide shows how to do that fairly after the setup. The Vision flow our lines use was built specifically to keep the tunnel cheap, and on a 5G line the mobile network, not the tunnel, is where the time goes. Setup first, then the measurement.

Copy the line

In the ProxyJolt dashboard, open My proxies and press the copy button on the VLESS / Xray row of your line. The vless:// string carries the server, port, your id and the REALITY settings; it is the credential, so keep it private and use Revoke access to reissue it if needed. VLESS runs on our server-based lines. Tick VLESS / Xray in the filters when buying or moving a line and the dashboard limits itself to locations that offer it.

If your link carries a flow value of xtls-rprx-vision, the client will use the direct path that skips a second layer of encryption for TLS traffic, which is where most of the tunnel's efficiency comes from. Every client in this guide reads that value from the link automatically.

Windows

Download the v2rayN With-Core zip, extract it and run v2rayN.exe.
Press Ctrl and V to import the line, right-click it and choose Set as active server.
Set System proxy to Set system proxy in the bottom bar, or leave it off and point a single app at SOCKS5 127.0.0.1 port 10808.
Right-click the line and run Test real delay to confirm the tunnel is up.

v2rayN releases

Mac, iPhone and Android

Mac and iPhone: install V2Box from the App Store, open Configs, tap the plus button, Import from clipboard, select the line and connect. Allow the VPN configuration.
Android: get v2rayNG, hit the plus button, choose Import config from clipboard, pick the line and tap V. Allow the VPN request.
On a phone, a quick check is to open a speed test site in the browser while connected and note the download figure for comparison later.

Linux and command line

Install Xray-core and write a config with a local SOCKS inbound and a VLESS REALITY outbound using the values from your link. Then run it and test with curl. The same curl commands are what we use for the fair comparison below.

xray run -c config.json
# through the tunnel
curl -o /dev/null -w '%{speed_download}\n' -x socks5h://127.0.0.1:1080 https://speed.cloudflare.com/__down?bytes=100000000
# through the plain SOCKS5 login on the same line
curl -o /dev/null -w '%{speed_download}\n' -x socks5h://USER:PASS@HOST:SOCKSPORT https://speed.cloudflare.com/__down?bytes=100000000

Xray-core releases

How to compare fairly

A single run through each path tells you almost nothing, because a 5G cell's throughput moves from minute to minute with the load on the tower. To compare the tunnel with the plain login, alternate them: tunnel, plain, tunnel, plain, five or more rounds each, within a few minutes, from the same machine, to the same test file. Then look at the middle values rather than the best or the worst. If the two sets overlap, the tunnel is not what limits you.

Use a large test file. Small downloads measure connection setup, and a tunnel adds one extra handshake at the start that a long transfer amortises to nothing. And compare against the plain login on the same line, not against your home connection, which is a different network entirely.

Repeat the whole exercise at a different hour before drawing a conclusion. A cell that is quiet at seven in the morning can be busy at nine in the evening, and a line that looked slower through the tunnel in one session may look identical in the next; the comparison you want is between the two paths, not between two times of day.

Same line, same minute, same file, alternating paths.
Middle value of several runs, not a single best run.
A file large enough that the handshake does not matter.
Rotation off during the test so the carrier IP does not change between runs.

If the tunnel is clearly slower

Check that the client is not routing through an extra hop such as a corporate VPN, that the device is not on battery saver, and that the flow in your link says xtls-rprx-vision rather than being empty; a missing flow makes the client fall back to a slower path. On phones, background restrictions on the client app can also cut throughput. If none of that explains it, send support both sets of numbers and the line name and we will look at the server side.

Keep the device on the same network for both halves of the test. Switching from Wi-Fi to mobile data, or from a wired desk to a laptop on the other side of the office, changes the first hop and swamps whatever difference the tunnel makes.

Frequently asked

Does the tunnel change the 5G speed of the line?

The carrier link is the same. The tunnel adds a handshake and a small amount of encryption work on a long transfer. Measure it with the alternating method above.

Which is better for a single speed-sensitive app, VLESS or SOCKS5?

Either works. Choose VLESS if you want the encrypted path or need to route an app with no proxy setting; choose SOCKS5 if the app already supports it and you want the least moving parts.

Can I run the speed check from the dashboard instead?

The dashboard's speed figure describes the line as our monitor sees it. Your own measurement through your own connection is the one that matters for your use.