Pillar guide13 min read·updated

Carbon vs Oxide in 2026: An Honest, Tested Comparison

Carbon and Oxide both compile C# plugins for Rust and share an API — but they differ on performance, update speed, and momentum. Here's an honest, tested comparison and a clear recommendation.

Most "Carbon vs Oxide" articles are generic feature rewrites, often outdated on how far Carbon has come. This is a current, practical comparison — what actually matters when you're choosing a framework for a server you have to keep running.

The short version

For a new server in 2026, install Carbon. It runs the same plugin ecosystem, performs better, and gets its post-wipe update out faster — which is the single thing that hurts most when it lags. Oxide is still perfectly usable and some long-running servers have no reason to switch, but the momentum is with Carbon.

What they have in common

It's easy to overstate the difference. Both:

  • Compile C# plugin .cs files at runtime — no precompiling.
  • Use the same hooks, the same [Info] attribute, the same permission system, and the same JSON config/data formats.
  • Run essentially the same library of community plugins.

Because Carbon implements the Oxide API, a plugin written for one almost always runs on the other — see how to make a Rust plugin for what that shared API looks like in code.

Where they differ

FactorCarbonOxide / uMod
PerformanceGenerally lighter; dynamic hook system only subscribes to hooks in useStable but heavier; all hooks always active
Post-update speedTypically updated within hours of a Facepunch updateSlower in recent cycles
Plugin compatibilityRuns Oxide plugins via API shim; rare internal-dependent ones failThe original target — everything written for Oxide
Hot-reload & toolingStrong file watcher, good console commands (c.*)Functional (o.*), less actively developed
MaturityNewer, but well past "experimental"Years of production hardening
MomentumWhere new development and most admin attention is goingIn maintenance mode; uMod site itself is largely static

The update-cadence point is the real one

Everything above matters, but in practice the deciding factor is this: Rust force wipes on the first Thursday of every month, and the update under it breaks plugins. Whichever framework gets itscompatible build out first is the framework whose servers are back online first. Lately that's consistently Carbon. A server that's down for fresh-map evening loses players that don't come back.

When Oxide still makes sense

  • An established server that's stable on Oxide, with an admin who knows it cold — "don't fix what isn't broken" is legitimate.
  • A plugin you depend on that explicitly only supports Oxide (rare, but check before switching).

Recommendation & next steps

New server: Carbon. Existing Oxide server that's working: consider switching at a wipe boundary, not mid-wipe, and only after confirming your critical plugins are green on Carbon (each plugin's page in our directory carries a tested verdict).

Ready to go? Install Carbon, then keep it current with how to update Carbon. The full server build is in the Rust server setup guide.

Frequently asked

Questions & answers

Is Carbon better than Oxide?
For a new server in 2026, Carbon is the pragmatic default — faster, updates quicker after force wipes, and runs the same plugins. Oxide still works and is battle-tested, but its update cadence has slowed. 'Better' depends on whether you value stability-of-the-known over performance and speed.
Can Carbon run Oxide plugins?
Yes. Carbon implements the Oxide API, so the large majority of Oxide plugins run unchanged — you just place them in carbon/plugins/. A small number that depend on Oxide internals won't, and the console tells you which.
Can I switch from Oxide to Carbon without losing data?
Largely yes. Plugin data and config files use the same JSON formats; you move plugins, configs, and data into the carbon/ folders. Back everything up first and verify each plugin loads after the switch.
Do I have to pick one?
Yes — run one framework at a time. Running Carbon and Oxide simultaneously conflicts. Pick based on this comparison and commit.
Ad Unit · inline