Hemmingway-1.5

Hemmingway-1.5

Writes everyday text such as emails, blog posts and product descriptions, and rewrites other AI text in a more personal voice. An open-weights model of about 27B parameters for non-commercial use.

Made by HemmingwayAI on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: Hemmingway-1.5 is an open-weights text generation model from HemmingwayAI, based on their earlier Hemmingway-1 and sized at about 27 billion parameters. It writes everyday pieces such as emails, blog posts, product descriptions, stories, speeches and comments. It can also take text from another AI and reword it so it reads as if a person wrote it.

Who it suits: Writers who want everyday text that sounds personal rather than stiff, and people who want an open model they can run on their own hardware.

Industry | Who uses it | How they use it | Example
Online shops | Sellers | Draft product descriptions | A short listing text for a handmade mug
Content and media | Bloggers and forum posters | Draft posts and comments | A blog post or a reply in a thread
Office work | Employees and managers | Write emails | A polite reply that declines an offer
Writing and speaking | Storytellers and speakers | Draft stories and speeches | A short speech for an event

Anyone who needs a model for paid commercial work can skip it, because the license covers non-commercial use only.

Where to use it:

• Python with the transformers library on Windows, macOS or Linux
• Your own computer, once the model files are downloaded
• The maker's web app, API and desktop and Android apps, which the card lists

Use case | Who | What they get back
Product descriptions | Shop owners | A short listing text with plain details
Email replies | Anyone with a full inbox | A polite reply in a personal tone
Rewriting AI text | Writers with a rough AI draft | The same text reworded in a person's voice

How to set it up:

• Python with pip
• Disk space for the weights, about 53.8 GB in BF16 or about 13.4 GB at 4-bit (an estimate)
• A machine that can run a 27B model with the transformers library

1. Install the Hugging Face hub tools.
2. Download the model files to your computer.
3. Follow the model card to load the model with transformers.

pip install -U huggingface_hub
hf download HemmingwayAI/Hemmingway-1.5

How to use it: Give the model an everyday writing ask, or paste a text and ask it to sound like a person wrote it. The card lists four thinking levels, starting at off and rising through low and medium to extra high. It reports about the same detector pass rate at each one.

Example: You ask for a customer review of a vacuum that quit after about a month. The card shows a reply of that kind, written like an ordinary customer. The brush starts to grind, the machine stops, and the writer gives three stars because it cleaned fine for that month.

For newcomers and experts: If you are new to this, try the maker's web app before you download a file of tens of gigabytes. Ask for a short product description or an email reply and read what comes back. The weights are large, so a download is the bigger step.

The model is a transformers release in Safetensors with BF16 weights of about 53.8 GB, based on Hemmingway-1. The card reports a 90 percent pass rate on AI-text detectors over 1,848 texts, against 86 percent for human-written text. Those runs are the maker's own, with one pass per detector and no retries. Sixteen quantized and one fine-tuned version of it exist.

You give an ask or paste an AI text
You pick a thinking level
The model writes or rewrites the text
You get the text back

Good to know:

• Detector results: the numbers come from the maker's own tests, not an outside review
• Size: the full weights are large, so plan for the disk space and memory

Title: Share of 80 everyday asks whose answers passed AI detectors, maker's own test
Hemmingway-1.5: 92%
Fable 5.1: 26%
GLM-5.3: 21%
GPT-6 Astra: 20%
Kimi K3: 18%

• Cost to run: no API key or paid service is named for the downloaded weights, only your own hardware
• Data it reaches: downloading fetches the model files to your computer, and the card names no other access
• Not for: commercial use under this license, and anywhere a school, employer or site asks you to disclose AI-written text

On Hugging Face, 13 November 2026

• Task: text generation
• Library: transformers
• Parameters: 26.9B
• Weights: about 53.8 GB in BF16; about 13.4 GB at 4-bit (estimate)
• Formats: Safetensors
• Downloads in the last 30 days: 3,322
• Likes: 61
• License: cc-by-nc-4.0
• Languages: en
• Based on: HemmingwayAI/Hemmingway-1
• Quantized versions: 18
• Fine-tunes: 1
• Tags: transformers, safetensors, qwen3_5_text, text-generation, qwen3.8, chat, creative-writing, humanizer, hemmingway, conversational

Links

• Model: https://huggingface.co/HemmingwayAI/Hemmingway-1.5
• Files: https://huggingface.co/HemmingwayAI/Hemmingway-1.5/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download HemmingwayAI/Hemmingway-1.5

Then follow the usage in its model card. Nothing in this kit runs by itself.