Dev Blog· 4 min
Why Apophis has its own runtime
We benchmarked a Rust engine against Java Minecraft with Sodium & co. The result: roughly 3 to 4 times the FPS and a 9 times faster startup.
The Apophis Runtime is a branded build of an MIT-licensed Rust engine for Minecraft 26.3. Before it went into the launcher, we measured whether it was worth it (phase R0).
The benchmark
Same world (seed 1234), same four flight paths, 1920×1080, VSync off, one run each. Machine: Core Ultra 7 265KF, 64 GB, RTX 5070 Ti. Opponent: Java 26.3 with Sodium, Lithium, FerriteCore and ImmediatelyFast.
| Apophis Runtime | Java + Sodium & co. | |
|---|---|---|
| Start to title screen | 1.7 s | 15.5 s |
| Start to in-world (render distance 16) | 3.9 s | 28.4 s |
| Avg FPS, render distance 16 (village) | 1561 | 413 |
| 1 % lows, render distance 16 (village) | 282 | 116 |
| Peak memory, render distance 16 | 1.2 GB | 3.0 GB |
What the numbers don't tell you
- One run per configuration on one machine – no confidence intervals.
- Both games rendered to an invisible target; with a visible window both are lower.
- The engine has 833 known gaps – where features are missing, it also does less work.
- Anything above ~1000 FPS is irrelevant to how the game feels. What matters are 1 % lows, load times, startup and memory.
Limits of the runtime
Minecraft 26.3 only, no Fabric or Forge mods, and on large servers deviating behaviour can trigger anti-cheat systems. There, Java mode with the companion mod is the safe choice – both modes sit side by side in the launcher.