---
title: New benchmark for POMA AI's document ingestion and chunking for RAG shows 77% token reduction ⬣ POMA AI
description: Our published benchmark: POMA AI cuts RAG token usage by 77% through document ingestion and hierarchical chunking.
canonical: https://www.poma-ai.com/featured/medium-benchmark-77-token-reduction
generated: Markdown variant of the page above, built from the prerendered HTML
---
Featured in Benchmark — Medium

# New benchmark for POMA AI's document ingestion and chunking for RAG shows 77% token reduction

Taking you to the original on **Medium** …

[If you're not redirected automatically, read it here →](https://medium.com/@POMA_AI/new-benchmark-for-poma-ais-document-ingestion-and-chunking-for-rag-shows-77-token-reduction-ac20c75da8e6)