zeroneverload/local-llm-benchmark repository overview

⁠Local LLM Benchmark

Reproducibly compare local and OpenAI-compatible LLMs with structured tests, clear metrics, and practical meta evaluation.

⁠Highlights

• Multi-system/model benchmark workflow
• Single tests, profile runs, and batch matrix mode
• Analysis, runs export, and live dashboard
• Meta evaluation (local judge + optional cloud exposure)
• Docker image for amd64 and arm64

⁠Quick Start

docker run -d --name local-llm-benchmark -p 3005:3005 -p 3015:3015 -v llm_benchmark_data:/app/data zeroneverload/local-llm-benchmark:latest

• 3005 = app
• 3015 = meta public port

⁠Docs

• Wiki (DE): https://github.com/zeroneverload/Local-LLM-Benchmark/wiki/Startseite-DE⁠
• Wiki (EN): https://github.com/zeroneverload/Local-LLM-Benchmark/wiki/Home-EN⁠
• FAQ (DE): https://github.com/zeroneverload/Local-LLM-Benchmark/wiki/FAQ-DE⁠
• FAQ (EN): https://github.com/zeroneverload/Local-LLM-Benchmark/wiki/FAQ-EN⁠