---
format: "aidr-story-markdown/v1"
id: "a346dbdc513519b517333a7cf5492dd507d9e983c32b0c1269eccb13bd5ff7bf"
canonical_url: "https://aidr.today/a346dbdc?lang=en"
title: "TufaLabs' tufazip Shrinks 1 GB of Wikipedia to 105.39 MB Losslessly"
lang: "en"
requested_lang: "en"
available_langs: ["en"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-19T22:34:06.000Z"
category: "Models"
topics: ["coding","llm","devtools"]
source_urls: ["https://huggingnews.com/ai/tufalabs-tufazip-shrinks-1-gb-of-wikipedia-to-10539-mb-losslessly-28a7858b","https://marketbrief.now/ai/tufalabs-tufazip-shrinks-1-gb-of-wikipedia-to-10539-mb-losslessly-28a7858b"]
summary: "TufaLabs released tufazip, a lossless compression tool that reduced 1 GB of Wikipedia text to 105.39 MB. The tool is the highest-ranking entry on the Large Text Compression Benchmark among models that use a neural network as the sole predictor. Other models that outscore tufazip use a combination of transformers and hand-built models."
---

# TufaLabs' tufazip Shrinks 1 GB of Wikipedia to 105\.39 MB Losslessly

> [Open the canonical story](<https://aidr.today/a346dbdc?lang=en>)

**Published:** 2026-11-19T22:34:06.000Z
**Category:** Models
**Topics:** coding, llm, devtools

## Summary

TufaLabs released tufazip, a lossless compression tool that reduced 1 GB of Wikipedia text to 105\.39 MB\. The tool is the highest\-ranking entry on the Large Text Compression Benchmark among models that use a neural network as the sole predictor\. Other models that outscore tufazip use a combination of transformers and hand\-built models\.

## Sources

- [Story source](<https://huggingnews.com/ai/tufalabs-tufazip-shrinks-1-gb-of-wikipedia-to-10539-mb-losslessly-28a7858b>)
- [Story source](<https://marketbrief.now/ai/tufalabs-tufazip-shrinks-1-gb-of-wikipedia-to-10539-mb-losslessly-28a7858b>)

