Llmspeed is a web‑based benchmarking utility that measures the inference speed of locally run large language models. By running the tool on existing hardware, users can obtain real‑time latency and throughput figures for various LLMs before deciding on an upgrade. The service helps developers and enthusiasts evaluate whether their current setup meets performance requirements for specific models.