I can be believe it’s faster than a poorly configured zsh (which unfortunately is quite common), but would be curious to see actual test results, with things like “input lag”, “time to first prompt”, “command lag”, etc. like is measured with the excellent zsh-bench (https://github.com/romkatv/zsh-bench#what-it-measures)
I've never tried zsh. I remember reading in the interwebz that if you go towards feature-parity with fish in a zsh setup, the latencies here and there, adds up and makes it slower than fish.
Here is a relevant blog post:
https://carlosbecker.com/posts/fish/