Instructions to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF",
	filename="QuickThinker-Qwen3.5-27B-Vision-Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": [
				{
					"type": "text",
					"text": "Describe this image in one sentence."
				},
				{
					"type": "image_url",
					"image_url": {
						"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
					}
				}
			]
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Use Docker

docker model run hf.co/pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Ollama
How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with Ollama:
```
ollama run hf.co/pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M
```

Unsloth Studio new

How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF to start chatting

Pi new

How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with Docker Model Runner:
```
docker model run hf.co/pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M
```

Lemonade

How to use pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.QuickThinker-Qwen3.5-27B-Vision-GGUF-Q4_K_M

List all available models

lemonade list

QuickThinker Qwen3.5-27B Vision GGUF

This release is part of the QuickThinker Series: a line of fine-tuned Qwen models focused on lower thinking-token usage, faster thinking, lower looping, cleaner stopping behavior, stronger prompt adherence, local tool chains, local tool use, and practical local use. These models are intentionally more direct and to the point.

QuickThinker Qwen3.5-27B Vision GGUF is based on Qwen/Qwen3.5-27B and packaged for local multimodal inference in GGUF format. The QuickThinker Series is built for local and quick inference, aiming to preserve the quality of the base models while keeping thinking enabled and still reducing the thinking-token budget by roughly 60 to 70 percent.

Base Model

This model is based on:

Qwen/Qwen3.5-27B

What This Release Tries To Improve

lower looping behavior
fewer wasted thinking tokens
stronger prompt adherence on structured tasks
better handling of contradictions, underdetermined prompts, and insufficient-information cases
more stable local assistant behavior
better tool calling for tools like OpenCode and Osaurus

Training Style

This release comes from the current final FineVine rebuild dataset, a custom curated dataset emphasizing:

concise but still substantive answers
cleaner stopping behavior
practical coding and reasoning tasks
image interpretation quality
political consistency grounding

The majority of the final training data is custom-curated and edited.

Included Files

This package currently contains:

QuickThinker-Qwen3.5-27B-Vision-Q4_K_M.gguf
QuickThinker-Qwen3.5-27B-Vision-Q6_K.gguf
QuickThinker-Qwen3.5-27B-Vision-Q8_0.gguf
QuickThinker-Qwen3.5-27B-Vision-f16.gguf
QuickThinker-Qwen3.5-27B-Vision-mmproj-f16.gguf

Important GGUF Note

For multimodal use in llama.cpp, you will typically need:

one language-model GGUF
the matching mmproj GGUF

Suggested Parameters

temperature = 0.6
top_p = 0.95
top_k = 20
min_p = 0.0
presence_penalty = 0.0
repetition_penalty = 1.0

Intended Use

This release is meant for:

local multimodal assistant use
direct-answer tasks
practical coding help
structured reasoning

Important Note

This model is not intended as a warm, roleplay-first chatbot. It is tuned more for directness, bounded reasoning, and practical usefulness.

This model is not a good fit for:

storytelling
role playing

Downloads last month: 86

GGUF

Model size

27B params

Architecture

qwen35

Hardware compatibility

4-bit

6-bit

8-bit

16-bit

Model tree for pt-ml/QuickThinker-Qwen3.5-27B-Vision-GGUF

Base model

Qwen/Qwen3.5-27B

Quantized

(197)

this model