> 1.16x faster than C++ and 1.12x than safe Rust

This alone is enough to tell you have no idea what you're talking about, and it's not even about the ambiguous usage of the word "faster".

There is no way for performance claims that are

a) this accurate,

b) this small,

to possibly be real.

You certainly just cherry-picked favorable benchmarks.

Ah, here it is: "Over sixteen benchmarks Goose runs at about...". First, let's make this statement actually correct: "Over sixteen microbenchmarks Goose runs at about...".

That's not the same statement as "1.16x faster than C++ and 1.12x than safe Rust". Not by a mile.

You need a test suite comprised of a wide variety of real world programs, not 16 microbenchmarks, possibly specifically chosen to play to the strengths of your language.