Instructions to use tinnlab/Qwen3.5-9B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tinnlab/Qwen3.5-9B-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="tinnlab/Qwen3.5-9B-GGUF")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("tinnlab/Qwen3.5-9B-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use tinnlab/Qwen3.5-9B-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf tinnlab/Qwen3.5-9B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf tinnlab/Qwen3.5-9B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf tinnlab/Qwen3.5-9B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf tinnlab/Qwen3.5-9B-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 tinnlab/Qwen3.5-9B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf tinnlab/Qwen3.5-9B-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 tinnlab/Qwen3.5-9B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf tinnlab/Qwen3.5-9B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/tinnlab/Qwen3.5-9B-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use tinnlab/Qwen3.5-9B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tinnlab/Qwen3.5-9B-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": "tinnlab/Qwen3.5-9B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tinnlab/Qwen3.5-9B-GGUF:Q4_K_M

SGLang

How to use tinnlab/Qwen3.5-9B-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "tinnlab/Qwen3.5-9B-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tinnlab/Qwen3.5-9B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "tinnlab/Qwen3.5-9B-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tinnlab/Qwen3.5-9B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use tinnlab/Qwen3.5-9B-GGUF with Ollama:
```
ollama run hf.co/tinnlab/Qwen3.5-9B-GGUF:Q4_K_M
```
Unsloth Desktop

How to use tinnlab/Qwen3.5-9B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf tinnlab/Qwen3.5-9B-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "tinnlab/Qwen3.5-9B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use tinnlab/Qwen3.5-9B-GGUF with Docker Model Runner:
```
docker model run hf.co/tinnlab/Qwen3.5-9B-GGUF:Q4_K_M
```

Lemonade

How to use tinnlab/Qwen3.5-9B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull tinnlab/Qwen3.5-9B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.5-9B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use tinnlab/Qwen3.5-9B-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf tinnlab/Qwen3.5-9B-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 tinnlab/Qwen3.5-9B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use tinnlab/Qwen3.5-9B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf tinnlab/Qwen3.5-9B-GGUF:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "tinnlab/Qwen3.5-9B-GGUF:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Qwen3.5-9B-GGUF (Q4_K_M) — Tin Nguyen Lab distribution mirror

This is a mirror. Tin Nguyen Lab did not train, fine-tune, or quantize this model.

This repository redistributes a single, unmodified GGUF file:


File	`Qwen3.5-9B-Q4_K_M.gguf`
Size	5,680,522,464 bytes (5.68 GB)
SHA256	`03b74727a860a56338e042c4420bb3f04b2fec5734175f4cb9fa853daf52b7e8`

Provenance

Role	Who
Original model / weights	`Qwen/Qwen3.5-9B` — the Qwen team
GGUF conversion + quantization	`unsloth`, in `unsloth/Qwen3.5-9B-GGUF`
This mirror	Tin Nguyen Lab — redistribution only, bytes unchanged

Mirrored from unsloth/Qwen3.5-9B-GGUF at commit 3885219b6810b007914f3a7950a8d1b469d598a5.

The SHA256 above is identical to the Git-LFS object id that unsloth/Qwen3.5-9B-GGUF publishes for the same file at that commit, so the file here is byte-for-byte the one distributed upstream. That equivalence is verifiable by anyone:

# what upstream publishes for this file
curl -s "https://huggingface.co/api/models/unsloth/Qwen3.5-9B-GGUF?blobs=true" \
  | python3 -c "import json,sys;print([s['lfs']['sha256'] for s in json.load(sys.stdin)['siblings'] if s['rfilename']=='Qwen3.5-9B-Q4_K_M.gguf'][0])"

# what this mirror publishes
curl -s "https://huggingface.co/api/models/tinnlab/Qwen3.5-9B-GGUF?blobs=true" \
  | python3 -c "import json,sys;print([s['lfs']['sha256'] for s in json.load(sys.stdin)['siblings'] if s['rfilename']=='Qwen3.5-9B-Q4_K_M.gguf'][0])"

Note on what this does and does not prove: the matching hash proves this copy is identical to unsloth's published file. It does not independently verify that unsloth's quantization is a faithful conversion of Qwen's original weights — that is a separate claim this mirror makes no attempt to establish.

Why this mirror exists

To keep a shipped "install a local model, no API key needed" flow working from a repository its maintainers control. It is a stability measure, not an improvement: there is nothing here that upstream does not have.

What is NOT mirrored

Upstream publishes 28 files (~148 GB): 25 quantizations plus multimodal projector (mmproj-*.gguf) and imatrix files. This mirror carries only Q4_K_M, so it serves text chat / tool use. It does not carry the vision projector and cannot be used for image or video input. For anything else, use unsloth/Qwen3.5-9B-GGUF directly.

README_upstream.md in this repository is unsloth's original model card, preserved verbatim from the mirrored commit, and is the authoritative documentation for the model itself (usage, chat template, quantization details).

Access

Public and anonymous — no token, no account, no gated agreement.

Licence

Apache-2.0, inherited from Qwen/Qwen3.5-9B. The full licence text as shipped by the original licensor is included as LICENSE; attribution notices are in NOTICE. Redistribution here relies on Apache-2.0 §4, and the licence and attribution are retained accordingly.

Downloads last month: 92

GGUF

Model size

9B params

Architecture

qwen35

Hardware compatibility

4-bit

Model tree for tinnlab/Qwen3.5-9B-GGUF

Base model

Qwen/Qwen3.5-9B-Base

Finetuned

Qwen/Qwen3.5-9B

Quantized

(476)

this model