It would be helpful to benchmark against other providers that sit atop MLX; this page tells me how OMLX does, but not why I should move from another (like LMStudio etc). I get that you have some features that you might only find in vllm, but how do I know that Ollama would be X tps slower? TBH not seeing competitors in a benchmark makes it less a benchmark and more a data sheet.
https://omlx.ai/benchmarks
