Original BASIC did not fit into tiny space, that is why while Dartmouth BASIC always compiled into machine code before execution, everyone that learnt BASIC in 8 bit systems thinks it was originally interpreted and compilers only came later, which was the compromise to make it fit into a few KB.

Jupiter ACE had its followers, and it was common to see ads on Your Sinclair and similar magazines for ROM replacements using Forth instead of BASIC.

That PDP-8 BASIC was a miracle of shoehorning as was everything else on the PDP-8.

My favorite minicomputer BASIC that I got to use was on RSTS/E on the PDP-11 which had split 64k address spaces for code and data and used fairly advanced compilation techniques. Roughly the RSTS/E experience was like having your own Apple ][ but with a hard drive and a little more oomph. I grew up in New Hampshire right next door to DEC's headquarters in Massachusetts and there were DEC minicomputers everywhere.

Microsoft had a compiled BASIC (like run a compiler, not compile interactively like Microware's BASIC09) on CP/M for the Z-80 which was a much better compiler target than the popular 6502.

I wrote a FORTH for the TRS-80 Color Computer using the OS-9 operating system which had maybe 2000-3000 lines of assembly code. FORTHs at the time often did block I/O directly to the disk but OS-9 had an API to access files that was pretty similar to Unix and my FORTH exposed that.