Show HN: Run MMLU benchmark on any LLM endpoint

I built it to quickly compare LLM hosting providers to weed out poor quantizations.