Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
smallevalsgithub.com · Low evidenceSmallevals is a lightweight local LLM evaluation framework that uses tiny 0.6B models to run entirely on CPU, GPU, or MPS. It evaluates RAG system retrieval quality by connecting to any vector database and generating synthetic question-answer pairs from document chunks.
Parse Score