How to Test Proxy Speed Before Your Team Commits
Before a proxy becomes part of a team's daily work, someone should confirm that it performs. Speed is really three measurements, not one: latency (how quickly the first byte arrives), throughput (how much data moves each second) and success rate (how reliably requests complete). A quick proxy that fails often wastes more of a team's time than a slower, steady one. This checklist shows how to measure all three and what the results mean.
Our proxy checker gives a quick first reading, and the manual methods below give a fuller benchmark. Only test proxies your team is authorised to use, and only against endpoints whose terms permit testing.
We write these guides for e-commerce, marketing and agency teams that pay for proxies by the gigabyte.
The three numbers to collect
- Latency / response time — how long one request takes to come back, normally measured in milliseconds. Lower is better. Distance and the proxy's own overhead account for most of it.
- Throughput / bandwidth — the sustained transfer rate, measured in Mbps. It decides how long large downloads and headless browsing take.
- Success rate — the share of requests that finish without an error. In day-to-day work it often matters more than raw speed.
Running the tests
A quick first check
Enter a proxy in the proxy checker: in one step it confirms the proxy works and shows the response time and exit IP.
Latency with cURL
Send a timed request through the proxy and note the breakdown your client prints: DNS lookup, connection, time to first byte and total time. Do this several times and keep the median, since any single sample can mislead.
Throughput
Fetch a test file of known, reasonable size through the proxy and divide its size by the download time. Pick a file from a host that permits you to download it.
Success rate
Send a batch of requests (100 works well in practice) and count the successful status codes. Repeat the batch at several times of day, so the figure reflects normal conditions.
What to expect from each proxy type
Expectations differ by type. Datacenter IPs respond more quickly than the others, while residential and mobile give up some speed in exchange for more trust. Treat the table as rough guidance rather than a promise.
| Proxy type | Latency | Throughput | Trust |
|---|---|---|---|
| Datacenter | Lowest | Highest (up to 1 Gbps) | Lower |
| Residential | Medium | Medium | High |
| Mobile (4G/5G) | Medium | Medium | Very high |
| Residential Premium | Low-medium | High | Highest |
Keeping the benchmark fair
A few principles keep the comparison honest:
- Test from the right region. Distance between the proxy and the test endpoint adds latency that has nothing to do with proxy quality.
- Use medians, not single samples. Network timings vary from one request to the next, so one slow result proves little.
- Benchmark the website without the proxy too. If the website itself is slow, the proxy is not to blame.
- Test the workload you will actually run — latency for many small requests, throughput for large transfers.
Our datacenter network runs on backbones of up to 1 Gbps with 99.9% uptime, so most of the variation your team sees in daily use comes from distance and the target rather than the proxy.
Next steps for your team
Benchmark a sample with the proxy checker, then compare tiers on the pricing page or buy proxies for a wider trial. Every plan supports HTTP/HTTPS and SOCKS5 with username/password or IP-whitelist authentication across 55+ countries, and you can check any proxy's status from the dashboard at any time.
Match the pool to the workflow
Every pool is billed per GB and run from the same dashboard, so one team can keep low-cost volume work and high-trust checks side by side.
High-volume collection from unprotected sites, automation, SEO tools
Market research, ad verification and geo-accurate checks on protected sites
Phone-first platforms, app and mobile ad testing
Hard targets where every failed request is costly
Covering several markets in one report? MIX packages combine multi-country residential traffic in a single order. Agencies buying on behalf of clients can apply for the reseller programme.
Get your team running
Per-GB pricing from $3.50 · 55+ countries · instant delivery · 30-day validity · no traffic logs · 24/7 human support
-
1
Size the job
Pick a proxy type in the proxy calculator and enter the gigabytes you expect to use to see the order total.
-
2
Order and pay
Choose the pool on the proxies page and pay by card through Stripe or in crypto through Cryptomus. Each order is billed per GB and does not renew on its own.
-
3
Share access
Credentials appear in the dashboard as soon as payment clears. Add them to each tool with a login or an IP whitelist, then send one test request before the real job starts.
Questions team leads ask
More playbooks for your team
What Are Proxies? A Primer for Business Teams
What a proxy server does for a business team: how the relay works, the main IP types and their per-GB prices,…
Read the playbook →Ad Verification Proxies: A Playbook for Ad-Ops Teams
How ad-ops teams check placements, creatives and landing pages from residential and mobile IPs per market, wh…
Read the playbook →Web Scraping Proxies: Picking a Pool for Each Data Job
The proxy pool a data team needs for each scraping target: datacenter for open sites, residential for protect…
Read the playbook →Datacenter vs Residential Proxies: Which Should Your Team Buy?
Decide workflow by workflow whether datacenter or residential proxies fit: cost, speed, trust and location ta…
Read the playbook →What Are Residential Proxies? When a Team Needs Home IPs
Residential proxies use real home IPs. Which team workflows need them, how a request travels, how they compar…
Read the playbook →Mobile vs Residential Proxies: Which Fits Your Team's Work?
Mobile proxies use carrier 4G/5G IPs; residential proxies use home broadband. Decide by platform which one yo…
Read the playbook →