Purpur vs Pufferfish: Performance Features, Config and Version Support

Published on | Updated

How Purpur and Pufferfish differ from Paper and from each other: what each fork changes, which config files matter, which Minecraft versions they support in 2026, and how to compare them on your own server.

Paper is the baseline for plugin servers. Purpur and Pufferfish both start from Paper and add something on top. Which one is faster for you depends on your world and plugins, so this post explains what each fork actually changes and how to find out which one wins on your server.

Comparing Minecraft server software performance

What Are These Forks?

The Minecraft server software family tree:

Vanilla -> CraftBukkit -> Spigot -> Paper -> Purpur
                                         -> Pufferfish

Paper adds its own chunk system, entity optimizations and a large set of configuration options over Spigot. Since 1.21 it also ships the spark profiler built in (PaperMC docs).

Purpur describes itself as "a drop-in replacement for Paper servers designed for configurability, new fun and exciting gameplay features, and performance" (Purpur on GitHub). Most of what you will notice is the extra configuration in purpur.yml: rideable mobs, farmland that cannot be trampled, villager tweaks and hundreds of other switches.

Pufferfish is "a highly optimized Paper fork designed for large servers" (Pufferfish on GitHub). It adds performance patches and a few server-owner features, and it keeps gameplay close to Paper.

Version support in 2026

Check this before anything else. In September 2026, Paper and Purpur both publish builds for Minecraft 26.3, though Paper's 26.3 builds are still on its alpha channel and 26.2 is its newest stable version (Paper downloads). Pufferfish's download page offers jars up to 1.21.8, and its public source code stops at 1.21.10 (the ver/1.21 branch on GitHub). If your server runs 26.1 or newer, Pufferfish is not an option, and moving a 26.x world back to a 1.21 server does not work: the PaperMC team is explicit that after upgrading a world to 26.1 "you cannot downgrade back to a lower version" (Paper 26.1 announcement). On a live server, that makes the stable 26.2 the Paper version to run until 26.3 builds leave the alpha channel.

Feature Comparison

PaperPurpurPufferfish
Based onSpigotPaperPaper
Runs Paper pluginsYesYesYes, per its README
Config filespaper-global.yml, paper-world-defaults.ymlPaper's, plus purpur.ymlPaper's, plus pufferfish.yml
Gameplay options (rideable mobs, trampling, mob behaviour)NoYes, manyNo
Dynamic Activation of Brain (DAB)NoNot in its docsYes
Villager lobotomize optionNoYesNot in its docs
Builds for Minecraft 26.x (September 2026)Yes (26.3 alpha, 26.2 stable)YesNo, newest jar is 1.21.8

The settings that make the difference

Pufferfish's best known feature is DAB, Dynamic Activation of Brain. According to the Pufferfish configuration docs, it "reduces the frequency of brain ticks" for entities that are farther from players. It applies to mobs that use the newer brain AI, such as villagers, axolotls, hoglins, zombified piglins and goats. You tune it with dab.start-distance (entities closer than this are left alone), dab.activation-dist-mod and dab.max-tick-freq, and you can exclude mobs with dab.blacklisted-entities. On a server with big villager trading halls, that is exactly where tick time tends to go.

Purpur's closest equivalent for villagers is lobotomize. With world-settings.default.mobs.villager.lobotomize.enabled: true, Purpur "Lobotomizes the villager if it cannot move (Does not disable trading)", per the Purpur configuration docs. Trading halls where villagers stand in one block cells are the typical use.

Both are worth knowing because they touch the same problem from two sides: lots of mobs with expensive AI, sitting near players who are not interacting with them.

How to compare them on your server

The result depends on your world, your plugins and what your players do, and you can find out in an evening:

  1. Back up the server, then make a copy of the world to test on.
  2. Run Paper on the copy with your normal plugins. With players online, start /spark profiler start --timeout 600 and note the median and 95th percentile MSPT from /spark tps.
  3. Stop the server, swap the jar for Purpur or Pufferfish (same Minecraft version), restore the same world copy, and repeat with a similar player load.
  4. Compare the two spark reports. Look at the entries that changed, not only the totals.

If the difference is small, pick the fork for its features and support, not for the performance.

When to Choose Each

Paper: The safe default. Maximum plugin compatibility, active development and the biggest community. Right for most servers.

Purpur: For servers that want Paper plus gameplay configuration: rideable mobs, custom enchant limits, crop and mob tweaks, and options like villager lobotomize. It tracks Paper's versions closely.

Pufferfish: Worth testing on a server that stays on Minecraft 1.21.8 or older (the newest jar on Pufferfish's download page) and has a lot of brain-AI mobs, where DAB can help. On 26.x it is not available.

Our Recommendation

For most servers, the choice between these forks matters less than view distance, simulation distance, mob farms and plugins. Start with Paper, or Purpur if you want its extra options, and let the spark profiler tell you where your tick time goes.

Paper and Purpur can be installed from the control panel on every Space-Node Minecraft hosting plan, and because Purpur is a drop-in replacement for Paper, switching between the two on the same Minecraft version keeps your world. If you want to run Pufferfish on a 1.21 server, ask us in live chat first.


Legal Notice

Legal Notice & Disclaimer: This article constitutes an independent, factual comparative review and critical analysis for educational purposes only. Space-Node is not affiliated with, endorsed by, or sponsored by any hosting provider mentioned herein. All brand names, logos, and trademarks referenced are the registered intellectual property of their respective owners and are used solely for identification and factual reference.

Fair Use & Review Rights: This review is protected commentary, comparison, and criticism. It is based on publicly available information, official pages where available, published documentation, and general hosting engineering analysis. Where hands-on testing is not explicitly stated in the article, no private benchmark or internal infrastructure access is implied. This constitutes lawful comparative review and criticism protected under fair use doctrine.

Factual Accuracy: Specific plan claims are based on public information available at the time of writing. Specifications, pricing, and service features can change, so readers should verify current details on the provider's official website before purchasing. We make no false or defamatory statements; criticism is limited to documented facts, clearly labeled opinion, or general hosting guidance.

No Consumer Confusion: This article makes clear that Space-Node offers distinct, independently-developed hosting infrastructure. We explicitly differentiate our services, pricing, and technical specifications. No reader could reasonably be confused about service provider identity.

Right to Comparative Advertising: Space-Node reserves the right to publish factual comparative information about competing services. This is a recognized right in consumer protection law and advertising standards. Accurate comparative reviews cannot constitute trademark violation, defamation, or unfair competition.

Limitation of Liability: Space-Node makes no warranty regarding third-party services reviewed. Readers are responsible for verifying information independently before purchasing. Space-Node disclaims liability for third-party service changes, outages, or policy modifications.

Space-Node Services: For Space-Node's own managed hosting solutions, visit Minecraft hosting or VPS hosting.