14 ms·

That mlx is for apple hardware only, though? Or did I misunderstand something.

• dragonwriter 6mo ago
  It needs a llama.cpp fork, too; so the stock runtime (based on stock llama.cpp) used by LM Studio presumably won't work for it.