Minecraft server statistics 2026: what 2,456 live servers look like

Most "Minecraft server statistics" pages repeat Hypixel's record and a total player figure nobody can check. This one is different: every player number below comes from our own pinger, which checks each server on this list every few minutes. A player count is what the server's own software reports when we ping it, and we check those counts for fakes before we use them.

The numbers below cover the 24 hours up to September 26, 2026 and 2,456 servers. Right now 2,937 servers on the list are online, with 410,244 players on them.

How many players does a normal Minecraft server have?

Twelve. For each server we took its typical player count over the day (the median of its hourly averages), and the median of those is 12. Half of public servers usually have fewer than 12 people on. The average is meaningless, because a handful of networks carry most players.

The top of the list is very concentrated. The 29 busiest servers hold half of every player we can see, and the top 100 hold 70%. The bottom half of the list, more than 1,200 servers, holds 1.4% between them. For an owner this is the important takeaway: you are not competing with Hypixel. You are competing with the 2,000-plus servers that have fewer than 100 players, and the difference between an empty lobby and a lively one is often ten people.

Java vs Bedrock vs crossplay

Java servers outnumber Bedrock ones 35 to 1 on the list. The Bedrock servers that are listed are mostly established networks, so their median says more about which Bedrock servers get listed than about Bedrock as a whole. The more useful comparison is inside Java.

We probe every Java server on the Bedrock port as well. The ones that answer, usually because they run Geyser, have a median of 26 players against 6 for Java-only servers. Some of that is because bigger networks are more likely to bother with Geyser, but the gap is large enough that a small Java server with an empty lobby should read how to make a crossplay server before it spends anything on advertising.

Account type matters less than owners assume. We test it ourselves by trying to join the way a player without a paid account would. Premium-only servers (1,148 of them) have a median of 8 players, and cracked servers (893) have a median of 18. Our fake-count checks flag both groups at about the same rate.

Which modes are crowded and which are open

Survival is listed on 946 servers, PvP on 535, SMP on 386. Those are the crowded shelves. Some smaller modes have far fewer servers but far more players per server.

A new Survival or SMP server is one of more than 1,100. A Crystal PvP server is one of 29, on a shelf where the typical server has 155 players. That does not mean everyone should open a Crystal PvP server, but if you are choosing what to build, the list tells you where demand outruns supply.

The number of modes a server lists also tracks its size. Servers listing one mode have a median of 9 players; four modes, 26; five modes, 42; six modes, 129. Networks are bigger, so this runs both ways, but if your server genuinely runs more than one mode, list them all: each one puts you on another list.

What busier listings have in common

These are correlations, not proof. Bigger servers publish more about themselves, and most listings on this site were set up by our team from each server's own public pages, so these columns mostly show how much a server puts out there. Still, the gaps are big, and every item in the table is free.

Only 2% of servers have a video on their listing, and they have the highest median in this table. Only 14% link a website. A complete listing also counts in our ranking: the listing strength checklist is part of every server's score, and the description is its biggest item. When we took this snapshot, the median listing strength was 3.5 and no server had reached the 8 needed for a Rising spot on page one, so the first owners to complete their listings will have those spots nearly to themselves.

• Addresses: 47% start with play., mc., join. or hub..
• Slots: the median server advertises 400 slots and fills about 5% of them. Servers with 20 or fewer slots have a median of 1 player; 21 to 100 slots, 5; 501 to 2,000 slots, 56. Setting a huge maximum doesn't make a server look busier: servers advertising over 2,000 slots have a lower median (41) than those with 501 to 2,000.
• Uptime: 93% of online servers answered every one of our pings since we started checking them. Our pinger went live on September 24, so that is two days of history, not a long record.
• Software: Velocity is the proxy of choice. 976 servers report Velocity in their version string, and 180 more run a Velocity fork, against 74 for BungeeCord and 23 for Waterfall. Paper (377), Purpur (84) and Folia (12) lead the single-server side.
• Fake counts: 58 servers (2.3%) carry a "player count looks unusual" flag right now, for counts that never move, exceed the maximum, sit exactly full at every check, or sit under a maximum that can't be right. We left them out of every player number on this page. They rank below honest servers of the same size.

How to use these numbers if you own a server

1. Benchmark honestly. If your typical count is 12, you are a normal server, not a failing one. At 30 you are in the top half of Survival, SMP, PvP and Minigames servers.
2. Fill the listing before you spend. Video, website, Discord, a 250-character description and every mode you genuinely run. It takes an evening, players check these before they join, and it's what our ranking rewards.
3. Add a Bedrock entrance. Java-only servers have a typical count of 6. Java servers that also let Bedrock players in have 26.
4. Pick a shelf with room. Read the mode table before you decide what your next season is.
5. Keep the count honest. A flagged count is marked on your page, loses some or all of its player points, and can't take a Rising spot.

Method

Every server on 9MinecraftServer is pinged from our own systems every few minutes using the standard server list ping for Java and the RakNet ping for Bedrock. For each server we took the median of its hourly average player counts over the 24 hours up to 9 pm UTC on September 26, 2026, counting only the hours it was online and only servers we had watched for at least 18 of those hours. We left out the 58 servers whose counts our checks flag as unusual. Listing details come from each server's listing on this site; most were set up by our team from the server's own public pages, and owners can claim and edit theirs. Medians are used everywhere because a few networks with tens of thousands of players make averages useless. We will refresh this page as the list grows and our history gets longer. The live figures at the top of this page and on the home page update automatically.

Quick questions