Hemmingway-1.5 GGUF

Hemmingway-1.5 GGUF

Imatrix GGUF quantizations of HemmingwayAI's Hemmingway-1.5, a 27B text model with MTP speculative decoding and a reasoning-effort prompt format. For people who run large models locally.

Made by bartowski 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: This repository holds GGUF quantizations of Hemmingway-1.5 from HemmingwayAI, a 27B text model. A third party made the files with llama.cpp and an importance matrix. The card lists text input and support for MTP speculative decoding. It does not say what the original model is meant for, so check the original model's page before you rely on it.

Who it suits: It suits people who run large local models and want a smaller file than the full weights. You choose a size to match your memory.

Industry | Who uses it | How they use it | Example
Local AI | Enthusiasts | Run a 27B model on a high-memory desktop | Chatting with the Q4_K_M file
Software | Developers | Test a large model's output and its tool-call layout | Trying the tool prompt format
Research | Model evaluators | Compare quant sizes of the same model | Measuring one size against another
Speed testing | Engineers | Try speculative decoding that uses MTP | Comparing speed with and without it

If you need commercial use, you can skip it, since the license is non-commercial.

Where to use it: You load a file in a tool that reads GGUF, built on llama.cpp.

• llama.cpp, which the files were made with
• Other GGUF-capable apps that support this model
• Windows, macOS or Linux with Python, to download the files

Use case | Who | What they get back
Local chat | Enthusiasts | Replies from a 27B model on your own machine
Tool calling | Developers | Function calls in the card's XML-style layout
Speculative decoding | Engineers | Faster generation through MTP, where the runtime supports it

How to set it up: You need enough memory for the file you choose and a runtime that supports the model. The repository holds many large files, so download one file rather than all of them.

• A GGUF-capable runtime
• A file that fits your memory, with Q4_K_M as the card's default pick
• Python, only for the Hugging Face download command

1. Pick a file from the table.
2. Download it from the repository.
3. Load it and use the prompt format below.

Title: Some files in the card
Quant | Size | Use
bf16 | 54.66 GB | Full weights, stored in split parts
Q8_0 | 29.12 GB | Top quality, more than most people need
Q6_K_L | 24.96 GB | Near-perfect quality, recommended
Q6_K | 23.86 GB | Near-perfect quality, recommended
Q5_K_M | 20.92 GB | High quality, recommended
Q4_K_M | 17.44 GB | Good balance, the card's default pick

pip install -U huggingface_hub
hf download bartowski/HemmingwayAI_Hemmingway-1.5-GGUF

How to use it: Wrap each chat in the layout below. The system block starts with a reasoning-effort line, and the assistant turn opens with a think tag. The card also shows a second layout for tool definitions.

<|im_start|>system
Reasoning effort is set to xhigh. Please think carefully through the task, validate key assumptions, consider plausible alternatives, and prioritize correctness, consistency, and clarity in the final answer.

{system_prompt}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
<think>

Example: You load the Q4_K_M file, fill in a short system prompt and ask for a short plan for a weekly meal prep. The model opens its think block, reasons through the request, and then writes the plan.

For newcomers and experts: A quantized file stores a model's numbers more coarsely to save space. Start with Q4_K_M, send one short prompt in the layout above, and see how it answers.

For experts, the files were built with llama.cpp release b11969 using an importance matrix. The card marks speculative decoding with MTP as supported and points to a section on it. The prompt format asks the model to reason at the highest effort level before answering.

The original model is converted to GGUF
An importance matrix guides the quantization
Several sizes are written, from bf16 to smaller
You download one file that fits your memory
You load it and send prompts in the layout

Good to know: The bf16 files come in split parts. Hugging Face estimates the 4-bit weights at about 13.7 GB, while the Q4_K_M file listed in the card is 17.44 GB.

Title: File size by quant type
bf16: 54.66 GB
Q8_0: 29.12 GB
Q6_K_L: 24.96 GB
Q6_K: 23.86 GB
Q5_K_M: 20.92 GB
Q4_K_M: 17.44 GB

• License: CC BY-NC 4.0, which is non-commercial
• Cost to run: none named
• Data it reaches: your prompts, run locally, and the files you download
• Not for: commercial use, or anyone who needs to know the model's purpose from this card alone

On Hugging Face, 14 November 2026

• Task: text generation
• Parameters: 27.3B
• Weights: about 13.7 GB at 4-bit (estimate)
• Formats: GGUF
• Downloads in the last 30 days: 13,596
• Likes: 17
• License: cc-by-nc-4.0
• Languages: en
• Based on: HemmingwayAI/Hemmingway-1.5
• Tags: gguf, qwen3.8, chat, creative-writing, humanizer, hemmingway, text-generation, imatrix, conversational

Links

• Model: https://huggingface.co/bartowski/HemmingwayAI_Hemmingway-1.5-GGUF
• Files: https://huggingface.co/bartowski/HemmingwayAI_Hemmingway-1.5-GGUF/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download bartowski/HemmingwayAI_Hemmingway-1.5-GGUF

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