That is too broad of a question for a too narrow of an answer. You can answer with broad statements and generalized estimations, but I don’t think they really answer the question.

Encoding video balances three things (extensible by two more):

• visual quality / equivalence
• size (stream/file size)
• encoding time
• decoding performance
• decoding feature set (compatibility)

The codec you use also has a high impact on compression ratio opportunities and capabilities. AV1, HEVC, AVC? 10-bit?

If we define that we do not care about encoding time, so we will use the very slow preset and use all codec features available, compression ratio and quality falloff still depends a lot on what you actually encode.

• Is it a cartoon with flat surfaces and mostly linear and partial linear or transformative movement? That can be compressed very well through differentials and transformation (movement).
• Is it a high-grain cartoon or movie? Fine, noisy details are hard to compress, they require more information.
• Does it have a lot of movement? A lot of vast movements and cuts? Less to keep and differentiate data with, so less efficient.

I suspect in higher resolutions the gaps between different visual data compression ratio differs more - because a difference is elevated through higher resolution/more data.

That being said, I don’t have or use 4K stuff, so I can’t even check for some rough numbers and visual content to size differences.

There is no “this much is very good enough for 4K movies” because the reasonable minimum very much depends on what the movie contains.