LLM inference speeds measured on Apple Silicon, raw data released

A benchmark study reports LLM inference speeds on Apple Silicon chips. The measurements cover various model sizes and configurations. Raw performance data is made publicly available. The dataset is released under a Creative Commons BY 4.0 license. Results compare inference latency across different hardware settings. The study aims to inform developers about Apple Silicon capabilities. The findings are hosted at the provided URL. Users can download the data for further analysis.