As far as I know, all three of those languages are interpreted. They might be good at slinging together predefined operations with fast implementations written in another language (C)… as long as you’re operating in parallel on large arrays, so that you spend more time inside the C functions than actually interpreting. But if you need to write your own operations or just do anything that’s isn’t massively parallel, none of those languages even compete. For example, you couldn’t write a performant compiler in them. In contrast, languages like Haskell and OCaml have compilers that generate native code – maybe not C-level native code, but still an order of magnitude faster than an interpreter.
Edit: For that matter, from quickly browsing the source code, it looks like Miranda is interpreted as well. So it’s absurd to say it’s faster than Haskell.
Most of the APLs don't automatically parallelise array operations either, because it's hard to know automatically when it's worth it. This is exacerbated by the fact that most APLs don't have a compiler, so the granularity of independent operations is fine.
While I don't know much about the mysteries of Ks implementation, I know that the most widely used industrial APL implementation, Dyalog, uses a pretty conventional explicit task parallel API for parallelism. They call it "isolates", and it's essentially about launching a thread that has its own internal APL state (with a lot of polish for convenience and communication, of course). There may be a few primitive operations that are automatically parallel internally, but they are rare.
Edit: For that matter, from quickly browsing the source code, it looks like Miranda is interpreted as well. So it’s absurd to say it’s faster than Haskell.