German Proxies in Numbers: Resi Core and Resi Prime Pools

Buyers often ask how many German addresses answer right now and how Resi Core and Resi Prime differ in practice. We ran both pools through our own benchmark with a Germany filter and collected the numbers: available pool size, providers, cities, response time, speed and spam list records. The country and plan overview lives on the German proxies page, and this article sticks to measurements.

How we measured

For each plan we took one line with the country parameter country-de, the same kind of line a customer gets in the panel. Through it, 100 parallel threads sent about 54,000 requests to an echo service that returns the exit address, plus 100 downloads of a 256 KB file to estimate speed. The requests came from our server, so your response time also depends on where your machine stands.

After the run we checked 2,000 addresses from each sample for geolocation, AS number and spam list records. We also sorted 1000 addresses per pool by city using the ip-api geolocation database.

Read the results with two caveats in mind. One line shows a lower bound: over a month of work with different lines you will see more unique addresses than in a single run. A short 256 KB file understates speed, because most of the time goes to TCP ramp-up. This setup works for comparing the two pools with each other, and it will not show the ceiling of the channel.

How many German IPs each pool has

Through one line, Resi Core showed 13,210 German addresses and Resi Prime showed 8,631. The Chao1 estimate landed close to these figures, and in the last tenth of responses new addresses made up 4% on Core and 1% on Prime, so both lines went through almost the whole slice available to them. The Prime curve reaches its plateau earlier. Home devices switch on and off, so at night German time you get fewer addresses than during the day and evening.

In practice, Core gives you a wider spread of addresses for rotation. Prime hands out a smaller set that you can narrow down to a state, a city or a provider.

The addresses are scattered across networks. Prime's 8,631 addresses sit in 5,833 /24 subnets and Core's 13,210 in 8,449, about one and a half addresses per /24. A site that counts logins by subnet sees points all over the country. The geolocation database placed 0.6% of Core addresses and 0.2% of Prime addresses outside Germany, single addresses in other countries.

Providers: whose lines these are

Four operators cover 80.3% of the Core sample and 83.3% of the Prime sample: Vodafone with its cable networks, Deutsche Telekom with DSL and fiber, Telefónica under the O2 brand and 1&1 Versatel. Regional networks follow: NetCologne in Cologne, EWE in the north, M-net in Bavaria, wilhelm.tel in Hamburg and Deutsche Glasfaser with fiber in smaller towns. The Core sample contained 114 autonomous systems and the Prime sample 88. You can check the exit address of any line yourself with the ASN lookup.

The pools differ in what else gets mixed in. In the Core sample 3.9% of addresses sit in hosting networks, in the Prime sample 0.1%. The geolocation database flagged 2.9% of Core addresses and 0.4% of Prime addresses as proxies. A scraper will not notice this mix. Accounts and ad cabinets are more sensitive to it, so we recommend Prime for them.

Cities: where the addresses live

Out of 1000 Core addresses the geolocation database found 370 different cities, and out of 1000 Prime addresses 322. Both pools share the top of the list: Munich, Frankfurt, Berlin, Düsseldorf, Hamburg and Stuttgart. Together these six cities give 30.8% of Core addresses and 35.2% of Prime addresses. The table shows addresses per 1000 next to each city's population.

Population: Destatis estimate for the end of 2021, rounded.

Frankfurt, home to 760,000 people, runs level with Berlin and its 3.7 million. In the Prime sample Düsseldorf and Stuttgart beat Hamburg, although Hamburg is three times the size of either. The likely explanation lies in the geolocation database: it ties some addresses of large operators to a network node, and the big nodes stand in Frankfurt, Düsseldorf and Munich. So the table shows which city the database names for an address. The router may stand in a neighboring town, and another database may name another city.

The east of the country is thinner in the samples: Leipzig and Dresden give under one percent each, Magdeburg and Halle about 0.7%. The Rhine-Ruhr area is dense: Düsseldorf, Essen, Dortmund, Wuppertal, Bochum, Duisburg and Gelsenkirchen together give more addresses than Berlin. Smaller towns such as Saarlouis and Diepholz also turn up in the list.

We tested city targeting separately on Resi Prime: Berlin, Munich, Hamburg, Frankfurt, Cologne, Stuttgart, Düsseldorf and Leipzig, four requests per city. All answers came from the requested city. Resi Core has no city choice, and rotation runs across the whole country.

Map: federal states and points of presence

Five federal states collect the most addresses: North Rhine-Westphalia thanks to the Ruhr and Cologne, Baden-Württemberg, Bavaria, Hesse and Berlin. Hamburg, Lower Saxony and Saxony come next. Cities in Schleswig-Holstein, Mecklenburg, Brandenburg and Thuringia did not reach the top 30. If you need a specific state, set it in a Resi Prime line and check the output with a test request.

Response time and speed

By the median, Prime answers four and a half times faster: 0.30 seconds against 1.37, and 93.7% of its requests fit within a second. Most Core requests fall between one and two seconds. Speed shows the same picture: 40.4% of Prime samples landed in the 10–25 Mbit/s range, and 71% of Core samples in the 1–5 Mbit/s range.

A scraper with hundreds of threads can live with a second of delay: it holds more connections, and a gigabyte of Core costs less. A browser in an antidetect tool, an ad cabinet or a script that waits for a response before its next step will do better on Prime: pages load faster, and you can set a session term of up to a day.

Spam list records

We checked 2,000 addresses from each pool against four public lists. Even if no address appears on two lists at once, at most 2.7% of Core addresses and at most 1% of Prime addresses are listed. Nearly all records belong to UCEPROTECT level 1. An address lands there automatically after hitting a spam trap and leaves after a week if no new hits follow.

We left the dronebl list out of the table, because we consider its records on home addresses false positives. If you run an address through the RBL check, that zone may show a mark. While you are at it, open the anonymity check to see what a site reads in your headers.

Which pool fits which task

• Scraping, rankings and prices in Germany. Resi Core: more addresses for rotation and a lower price per gigabyte. You set the country and a session term from 3 minutes to 24 hours in the line.
• Accounts, antidetect and ad cabinets. Resi Prime: fast response, 0.1% hosting addresses in the sample, a session term anywhere between 3 minutes and 24 hours, plus a choice of state, city and provider.
• Thousands of threads and traffic without a meter. Unlimited packages: a European pool with German addresses in an ip:port list, paid by term.

All per-GB plans draw on one balance, and purchased traffic stays valid for three months. If your software expects another line format, run the list through the proxy converter.

How to repeat the test

Create a list with a Germany filter in the panel, take the login, password, host and port of a SOCKS5 line and start the loop. It prints the exit address and response time of each request:

for i in $(seq 1 20); do
  curl -s --max-time 15 -w " %{time_total}s\n" \
    --socks5-hostname "LOGIN:PASSWORD@HOST:PORT" https://api.ipify.org
done

You can run the list of addresses through the SLP checker or the Linux script on the same page, which checks liveness and spam lists. The free German proxies help you debug a script before buying, and the full open list lives on the free proxies page. Those addresses last for hours, and many of them sit in hosting networks.

If your numbers for Germany differ noticeably from these, write to support with the exit addresses and request times, and we will take a look.