goal : build a game - 1st solution : use framework + n days dev time - 2nd solution : use libraries + 10xn days dev time

That's true in most cases, if you want to make a game as fast as possible, if it's not too ambitious, if you like the framework, and if the constraints are satisfying.

Frameworks are not so cut and dry compared to libraries. Not using a framework is not necessarily a bad idea.

Godot lets you script with native code (GDNative) and more importantly lets you access the very low level parts of the engine should you so wish, allowing you to implement advanced custom functionality performantly.

I do get your point but honestly unless you're a big studio or a custom engine enthusiast, there's very few logical reasons not to use a framework.

Exactly. Also, most of the time frameworks/Engines has the tool you need to build a game. Only in some specific technical bottle neck you need to ditch game engines and go for libraries or some in-house tools. But in that case, it would be pretty artist-unfriendly.