Free tool
FiveM RAM + slot calculator
How much RAM does my FiveM server need? Pick your framework, expected player count, and how heavy your resource setup is — we'll recommend a RAM tier sized for your Hetzner cpx instance type with headroom for peak hours.
Peak concurrent during your busiest hour, not total members.
Recommended tier
24 players × 200 MB/slot × 1.15 (medium resources) + 1.5 GB baseline ≈ 6.9 GB needed. Pick: 8 GB tier (cpx32).
Get this tier on CoalHosting →How we size the recommendation
FiveM is CPU-bound: single-thread perf matters more than total RAM. The recommendation maps to Hetzner's cpx-line instance types (the same hardware most serious RP communities self-host on):
- cpx22 — 3 vCPU, 6 GB RAM, ~24 slot ceiling, €15/mo
- cpx32 — 4 vCPU, 8 GB RAM, ~32 slot ceiling, €22/mo
- cpx42 — 8 vCPU, 12-16 GB RAM, ~48-64 slot ceiling, €35-49/mo
ESX and QBCore each add roughly 200-400 MB of resident RAM and modest CPU per tick. Heavy jobs / AI / map resources stack on top of that and can push you up a tier.
Want this tier on CoalHosting?
Every tier above maps 1-to-1 to a CoalHosting FiveM plan. EU only (Helsinki), txAdmin + ESX/QBCore one-click, daily backups (with the MariaDB dump), and a 30-day no-questions-asked refund during the soft launch.
FAQ
How much RAM does a FiveM server need?
FiveM is more CPU-bound than RAM-bound. 6 GB handles small private RP (<24 players, vanilla + a handful of scripts). 8 GB is the sweet spot for an ESX or QBCore base server with 24-32 active players. 12 GB only becomes necessary once you have 30+ concurrent players with heavy framework + jobs + AI resources. 16 GB is for 50+ active RP at peak hours.
How many slots can my server handle?
On the cpx-line shared-vCPU plans we use, expect ~32 slots on cpx22 (3 vCPU), ~48 on cpx32 (4 vCPU), and ~64 on cpx42 (8 vCPU) before TPS drops below the smooth RP threshold. The calculator below uses these limits as the slot recommendation per RAM tier.
Does framework choice (ESX vs QBCore vs vanilla) affect the recommendation?
Yes, mildly. Vanilla + light scripts is the lightest. ESX Legacy and QBCore both add similar overhead (a few hundred MB of RAM, modest CPU per tick). Heavy job + AI resources on top of either framework can push you up a tier.
Is the calculator a hard rule or a starting point?
Starting point. Real RAM use depends on which resources you install, how many AI peds spawn, and how aggressive your map / mission resources are. Pick the tier the calculator recommends; if you're seeing memory pressure after a week, upgrade. CoalHosting's dashboard shows live RAM use so you can decide based on real data.