Show HN: RigMark benchmarks local AI the way coding agents use it

Category: ai-ml

Tags: benchmarking, local-ai, llm-serving, performance-testing, openai-compatible

Score: 6.3/10 (Innovation: 6, Technical: 6, Documentation: 7, Utility: 6)

RigMark is a reproducible serving benchmark for local AI models, simulating coding-agent workloads with prose, code, structured output, and prefill/concurrency tests. It stands out for its emphasis on traceable receipts, cache-busting prefill measurement, and a model-agnostic approach across OpenAI-compatible stacks, making local AI performance claims more comparable and verifiable.

Target audience: AI engineers, local AI enthusiasts, hardware benchmarkers, and infrastructure teams evaluating LLM serving stacks

Repository: https://github.com/alexellis/rigmark · Python · MIT · 25 stars

View on Hacker News