A solo 2D game built mostly through scripts and a content-heavy 3D game built by several people put very different pressure on an engine. “Can support larger projects” is too broad by itself. The real questions are how many assets you expect, whether multiple people will edit the project, and what platforms and visual quality you actually need.
For a beginner who wants room to grow, Unity is probably the least risky compromise. C# has a conventional programming model, the engine covers 2D and 3D reasonably well, and there is enough documentation and third-party material to get unstuck. The downside is complexity that does not initially look like complexity. Render pipelines, package versions, input systems, editor settings, and plugins can turn a simple tutorial into a configuration exercise. Unity is beginner-friendly in the sense that answers are easy to find, not necessarily because the engine itself is simple.
Godot removes a lot of that friction. Projects are small, iteration is quick, and scenes and scripts are generally easy to inspect. Another practical advantage is that many project files are text-based, which makes source control and reviewing changes less painful. That matters once a project grows or another person joins. Godot can absolutely handle substantial games, but its ecosystem gives you fewer ready-made solutions when you hit specialized requirements. You may spend less time learning the editor early on and more time implementing missing project-specific tools later.
I would push back slightly on using “large 3D project” as an automatic reason to choose Unity. If large means realistic environments, complex animation, cinematic tooling, or a lot of high-detail assets, Unreal may be the more appropriate foundation despite its steeper start. Blueprints can get gameplay running quickly, but they do not make the rest of Unreal small or simple. Build times, storage use, binary assets, source control, and the C++ layer become real concerns. It is a poor default for someone making a basic platformer, but a reasonable first engine when the intended game genuinely depends on Unreal’s strengths.
There is another trap here: assuming your first learning project must become your future large project. It usually should not. Beginner code accumulates structural mistakes regardless of engine, and carrying those mistakes forward is often worse than rebuilding a prototype after you understand the problem. Pick an engine for the next small game, not the hypothetical game you might make three years from now.
My practical choice would be Godot if you are still learning programming and game logic, Unity if you already know you want broad platform support and a larger ecosystem, and Unreal only when high-end 3D is a firm requirement rather than a vague ambition. Spend a few evenings making the same tiny mechanic in your top two choices. Include input, UI, saving, and an exported build. The engine that makes those ordinary tasks understandable is a better choice than the one producing the nicest editor screenshot.
