Video summary

I Made The Same Game in Unity and Godot, Which Engine is Better?

Main summary

Key takeaways

Technology

Overview

Thomas Grové (game designer) recreated the same simple horror-game prototype in Unity and Godot to compare performance, usability, workflow friction, build/load times, install size, and profiling results. He aimed for an apples-to-apples comparison by starting from his Unity work (already had the character controller, camera transitions, scene transitions, and a custom dither shader) and porting it to Godot.


Key technological comparisons & findings

1) Install size / disk footprint

  • Unity install size: about 21 GB (Unity Hub + editor)
  • Godot install size: about 164 MB

Takeaway: Godot is ~129× smaller, which can matter on limited-storage machines.


2) Editor + project startup time

From “nothing” to loading his project:

  • Godot: ~13.5 seconds
  • Unity: ~1 minute 20 seconds

Takeaway: Godot starts much faster for initial editor/project load.


3) Gameplay parity / visual matching

  • He ran both versions simultaneously with one controller controlling both.
  • He couldn’t reach a perfect one-to-one match (e.g., walk speed tuning differed, character/exit timing mismatched).
  • Visuals:
    • Dither effect and basic lighting/textures looked “fine” in both.
    • He preferred Godot’s look slightly, especially corner fading to black applied more liberally in Godot.

4) Camera follow system

  • Unity approach: Cinema Machine
  • Godot approach: custom follow camera

Takeaway: Though he was hesitant to roll his own, he preferred his Godot solution for controllability/simplicity.


5) Workflow / architecture differences

Overall he spent about similar time on both, but:

  • Even though Godot might feel like it should be faster after doing Unity first, he had less motivation to redo work.
  • Learning Godot nuances was sometimes frustrating for someone experienced with Unity.

Example friction points:

  • Rigged characters / animation requirements
    • Unity: can work with T-pose or A-pose; root motion optional
    • Godot: prefers T-pose
      • For non-root-motion, root motion should be disabled, which may require re-export or conversion steps (e.g., Mixamo versions without root motion, Blender edits, or a plugin)
    • Unity FBX workflow is simpler; Godot requires extra steps to save/map animations into resources/animation libraries.

6) Rendering pipeline maturity & custom shader needs

  • He felt Unity URP is more mature than Godot’s pipeline for his requirements.
  • Shader requirement:
    • He needed a dither effect applied after transparency, not before.
    • In Unity, he created a custom compositor/HLSL shader to place the effect correctly.
  • He believes this may reduce performance a bit (later reflected in metrics).

7) Editor UX “gotchas”

  • Editor behaviors can hide properties until you expand them (example: clicking to reveal material properties).
  • Some workflows require material variants/instances so edits don’t apply globally.

8) Script/shader compile times (Unity pain point)

Unity compile time for a major script (character controller) after changes:

  • Godot: ~0.5 seconds
  • Unity: ~15.5 seconds

Takeaway: Unity compilation was ~31× slower in this test, contradicting his past experience with Unity’s “instant” iteration.


9) Export / build size and build times

  • Unity seemed to export “optimized” assets by default; Godot can do similar but is not on by default.
  • First Godot export produced a larger file because source assets were included.
    • He fixed this by reorganizing assets and excluding source folders/resources during export.

Build time to PC (first build):

  • Unity: ~15 minutes
  • Godot: ~2–2.5 seconds

Second Unity export:

  • ~52–53 seconds (still far slower than Godot, but improved vs first run)

10) Game launch time from builds

After exporting, load times:

  • Godot: ~2 seconds + frames
  • Unity: ~10.5 seconds

Takeaway: ~5× faster launch for Godot in his test.


Performance / FPS & profiling results

FPS target and VSync

  • Intended target: 60 FPS
  • He disabled VSync to measure “headroom.”

Initial profiling confusion (Godot shows different profilers)

  • Godot “appeared” to run at ~950 FPS while Unity was ~200–300 FPS.
  • Cause: Godot has two profilers (logic/physics vs rendering).
  • He initially looked at the wrong one; after correcting, the rendering comparison became more meaningful.

Frame rate (final export, barebones room)

  • Unity final export: ~650 to ~730 FPS
  • Godot final export: similar range in the corrected measurement
  • He notes results were close, except for one case.

GPU vs CPU utilization

  • Godot: ~95% GPU, ~8% CPU → more GPU-bound
  • Unity: ~76% GPU, ~29% CPU → more CPU overhead

Lighting change test in Godot

  • He changed global illumination from Voxel GI to Lightmap GI and adjusted resolution.
  • Result: Godot FPS increased to about 630 FPS.

System load restart test (reducing background interference)

He restarted and reran both without background apps (Godot/Unity/DaVinci Resolve/OBS):

  • Godot: ~1050 FPS
  • Unity: ~850 FPS
  • He notes results are somewhat inconclusive due to test conditions.

Overall conclusion

  • Based on his collected metrics, Godot beat Unity on every metric except the final output FPS, and even there the FPS results were pretty close.
  • Both engines exceeded his practical needs by a large margin (he reported ~600 FPS headroom), so performance was not the deciding factor.
  • He expects to continue with Godot due to:
    • faster iteration/build/load
    • more tailored/custom feel
    • generally strong performance

Mentions of guides/tutorial-style points

  • Editor/property visibility gotcha: properties can be hidden until you click/expand them (relevant when following tutorials).
  • Godot export gotcha: you must configure resource inclusion/exclusion manually to avoid packaging unnecessary source assets (Unity-like behavior isn’t automatic by default).
  • Godot animation pipeline gotcha: pose/root-motion expectations and extra steps/resources mapping.

Main speakers / sources

  • Thomas Grové (primary speaker; game designer conducting the Unity vs Godot test)

Original video