Arma Reforger Server Hosting Requirements: How Much RAM?

There is no official RAM-per-player requirement for Arma Reforger. Its server docs publish a real memory reading and a Docker example, but neither proves what a particular scenario needs. For a managed server, we recommend Game Hosting’s Arma Reforger plans: the 8, 12 and 16 GB tiers state the allocation directly, while each keeps the same Ryzen 9 9900X and four CPU threads. Treat their player bands as starting recommendations, not capacity guarantees. This guide shows how to size a budget, read the available evidence and check your own workload before moving up.

Arma Reforger server hosting requirements: set a RAM budget#

Start with the kind of session you run, then choose an allocation you can check. A small co-op group with few mods is a lighter starting point than a large Conflict community or a custom Game Master session filled with AI. That distinction is useful; an exact number of gigabytes per player is not available from Bohemia.

Game Hosting publishes three Arma Reforger memory tiers. Its plan page recommends about 16 players on 8 GB, 32 on 12 GB and 64 on 16 GB. Those are the service’s population recommendations, not player-slot limits, a guarantee that every scenario will fit, or the result of a controlled benchmark in this guide. The plans differ by memory; CPU hardware is the same across the three.

For readers weighing a rented plan against a self-hosted server or another provider, the Arma Reforger hosting checklist covers the rest of that decision. Here, use the population bands only as a first budget: if your group runs a different scenario, high AI counts or a substantial mod collection, verify the real workload instead of treating the nearest band as proof.

Read Bohemia’s real log without turning it into a minimum#

Bohemia’s -logStats documentation includes one actual server output: 2 players, 104 AI and Mem: 3291106 kB. The same line reports FPS, frame time, vehicles and replication counts. It is a useful example of what the server can measure, not a minimum requirement or a prediction for a full server.

The number comes from the game’s internal estimate. Bohemia warns that its Mem figure does not match operating-system memory measurements exactly. A panel showing container memory can therefore report a different number from -logStats at the same moment. Compare like with like, and do not subtract this sample from a plan’s allocation to claim how many players “fit.”

The example also combines a specific player count with a large AI count. It does not say which scenario produced it, how long the server had been running, what mods were loaded or what the operating system used. Those missing details are why one real log is still only a snapshot. The best use of the sample is to recognize the fields and understand their limits.

See Bohemia’s startup parameter reference for the exact output and the -logStats interval. Do not use a lone reading to promise stable performance at 32 or 64 players.

Treat the 6 GB container example as an example#

Bohemia’s server-hosting guide shows a Docker configuration with Memory: 6 GB and CPU: 4 cores. That is a published example environment. The page does not call it a minimum, compare it against each scenario or say it sustains a particular player count.

A container’s memory limit is also not the same thing as the game’s internal Mem number. The container has an allocation and its operating system uses some of that memory; the server reports its own estimate. If a provider gives you only a slot count, ask for the actual memory allocation. If it gives you an operating-system or container graph, note which measure it shows before comparing it with server logs.

The example is helpful as a reality check: the official documentation describes a server deployment with a fixed allocation instead of declaring a universal number for every group. It does not prove that 6 GB is enough for your mission, nor that a 12 GB plan will perform twice as well as a 6 GB one. CPU allocation, server version, scenario and add-ons still matter.

Account for AI, view distance and mods#

Player count is one input, not the whole workload. Record the number of AI entities as well as players. Also note the scenario, mod list, server view distance, network view distance and how long the session has been running. Those are configurable parts of a Reforger server; there is no official formula that converts them into a fixed RAM amount.

The server configuration reference documents settings such as operating.aiLimit, gameProperties.serverMaxViewDistance, gameProperties.networkViewDistance and game.mods. Their allowed values describe configuration, not guaranteed memory savings or costs. Avoid claims like “each mod needs X GB” unless the mod’s author or a controlled measurement supports that exact statement.

A mod pack can change more than memory: dependencies, version compatibility and console storage can affect whether the group can use it at all. Test with the same mod versions and scenario you intend to keep. If you name or recommend a specific mod, check its current Workshop page and credit its author; this sizing guide does not need a “best mods” list.

Choose a starting tier for your group’s workload#

Use the table as an initial plan map, not as a technical requirement chart. The population figures are Game Hosting’s current recommendations; neither they nor the workload descriptions below are benchmark results.

A four-player Combat Ops group is below the mode’s official player ceiling, but that does not establish its memory use. A 10-player Game Master setup can be busier than its headcount suggests if the Game Master places many AI entities. Conflict can have a larger population, yet two sessions with the same player count may still differ because of AI, mods and session settings. There is no published per-mode RAM chart that settles those cases.

If you need one place to start, match the plan’s memory allocation to your expected group, then leave enough room to test the actual setup. Move to a higher tier when the logs or host’s operating-system graph show sustained pressure under your normal busy load. A plan recommendation is a useful estimate for choosing a starting point; it cannot replace a measurement of your exact workload.

Check the running server before moving up#

A Reforger server can write performance statistics with -logStats. To log them every 10 seconds, Bohemia documents this launch argument:

-logStats 10000

Collect readings during a representative session rather than checking only at startup. Alongside each sample, record the scenario, version, players, AI, vehicles, mods, view-distance settings and session duration. Compare the server’s Mem estimate with the host’s operating-system or container measurement separately; they are different metrics.

If you rent a server and cannot change startup arguments, ask the host which memory graph is available and whether it can provide the server log. Look for a repeated pattern during the busy part of your own session, not one isolated high or low value. If memory keeps approaching the allocation or the server reports memory-related failures, save the relevant logs and ask the host whether more memory is appropriate before changing settings or wiping anything.

This is an observation method, not a promise that a single threshold fixes every slowdown. Low server FPS can have a different cause, so include the FPS and frame-time fields in your notes rather than treating RAM as the answer to every performance problem.

Keep stable and Experimental readings separate#

As checked on 11 October 2026, Bohemia’s news lists stable version 1.8.0.13 and Experimental version 1.9.0.10. The 1.9.0.10 post says the update went to separate Steam and Xbox Experimental applications. Experimental and the main version are separate builds, so label every reading with its branch and version.

Do not use a log from Experimental to size a stable server, or combine readings taken before and after a game update as if they came from one workload. The scenario, mods and settings need to match too. For an update that changes vehicles, AI or server performance, take a fresh baseline with the build your group actually runs.

The official Arma Reforger news page is the place to check current releases before comparing a new reading with an old one. If the measured load no longer resembles your current server, repeat the observation before buying memory based on the outdated sample.

Frequently asked questions#

How much RAM do you need for an Arma Reforger dedicated server? There is no universal official minimum. Start with a stated allocation, then compare readings from your scenario, player count, AI and mod list.

Is 8 GB enough for a 32-player Arma Reforger server? Game Hosting positions 8 GB at about 16 recommended players and 12 GB at about 32. That is plan guidance, not a hard cap or a performance guarantee; measure your workload before treating it as a fit.

Does Game Master need more RAM than Conflict? No official RAM-by-mode chart answers that. AI and mod load can change the work, so compare representative sessions rather than assigning a fixed amount based on mode name alone.

If your next decision is plan selection, compare the current Arma Reforger hosting options with the complete host-selection checklist. Our Arma Reforger monthly cost guide shows dated prices and bill splits without treating a price as a capacity test. Keep the log beside the plan you choose so you can review the same workload after a change.