# llm-benchmark

A cross-platform command-line tool that measures the real tokens-per-second throughput of local LLMs running through Ollama.

llm-benchmark measures how fast local LLMs actually run through Ollama, reporting tokens-per-second throughput from a simple, cross-platform command-line tool. Depending on the amount of RAM detected on the host, it automatically pulls an appropriate set of Ollama models to benchmark, ranging from small models on systems with limited memory up to larger models on systems with much more RAM.

It requires a working Ollama installation and runs inside a Python virtual environment, tested on Python 3.10 and above.

It installs with pip, pipx, or uv.

Website: https://llm.aidatatools.com/