Metrillm

MetriLLM is an open-source benchmarking tool for evaluating local large language models (LLMs) on your own hardware. It provides quantitative metrics like speed, quality, and resource usage, aggregates a global score, and enables easy sharing of results to a public leaderboard. Ideal for developers, AI enthusiasts, and engineers working with self-hosted or local LLMs who want to quickly compare model performance and suitability across different setups.