# Open-source Humanizer model claims 95% human-written rewrite rate, but tests show mixed results

Digest AI · Marketing & Small Business · published 2026-10-28T12:00:00Z

Canonical: https://digestai.news/story/open-source-humanizer-model-claims-95-human-written-rewrite-rate-but-t

## Summary

A free, open-source AI model called Humanizer—fine-tuned on 28,598 AI-to-human text pairs—claims to rewrite AI-generated content into human-like writing with 95% success, per its developer. The 12-billion-parameter model, built on Google’s Gemma 4 12B, has been downloaded over 143,000 times on Hugging Face and supports local deployment via tools like `llama.cpp` or `Ollama`. Its developer reports that 95% of 210 English rewrites passed Originality.ai’s human-written filter, though this was not independently verified. The model avoids AI detectors in training, focusing instead on factual accuracy and natural phrasing.

The author noted inconsistency in outcomes and questioned detector reliability. Humanizer runs offline on Windows, macOS, or Linux with GPU support, offering an alternative to paid cloud-based humanizers. The project’s GitHub and Hugging Face pages provide setup guides for non-developers, though advanced users can fine-tune parameters like context window and Flash Attention for performance.

## Key points

- Humanizer, a 12B-parameter open-source model, claims 95% of its rewrites pass Originality.ai’s human-written filter, but tests show inconsistent results
- Developed on Google’s Gemma 4 12B, it avoids AI detectors in training, focusing on factual accuracy and natural phrasing instead

## Why it matters

Humanizer offers a free, locally run alternative to paid AI humanizers, appealing to educators, writers, and businesses avoiding cloud dependencies. Its inconsistency highlights the limits of detector scores and the need for manual review, but its open-source nature could spur further research into human-like text generation.

## Sources

1. [Can This Open-Source Model Beat AI Detectors? Let’s Test It Locally](https://kdnuggets.com/can-this-open-source-model-beat-ai-detectors-lets-test-it-locally) (KDnuggets, 2026-10-28)

## Cite

Digest AI, "Open-source Humanizer model claims 95% human-written rewrite rate, but tests show mixed results", 28 October 2026, https://digestai.news/story/open-source-humanizer-model-claims-95-human-written-rewrite-rate-but-t

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/open-source-humanizer-model-claims-95-human-written-rewrite-rate-but-t.json