Instructions to use Lucebox/Qwen3.6-27B-DFlash-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Lucebox/Qwen3.6-27B-DFlash-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Lucebox/Qwen3.6-27B-DFlash-GGUF",
	filename="dflash-draft-3.6-q8_0.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use Lucebox/Qwen3.6-27B-DFlash-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0
# Run inference directly in the terminal:
llama-cli -hf Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0
# Run inference directly in the terminal:
llama-cli -hf Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0

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 Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0
# Run inference directly in the terminal:
./llama-cli -hf Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0

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 Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0

Use Docker

docker model run hf.co/Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0

LM Studio
Jan
Ollama
How to use Lucebox/Qwen3.6-27B-DFlash-GGUF with Ollama:
```
ollama run hf.co/Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0
```

Unsloth Studio new

How to use Lucebox/Qwen3.6-27B-DFlash-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 Lucebox/Qwen3.6-27B-DFlash-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 Lucebox/Qwen3.6-27B-DFlash-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Lucebox/Qwen3.6-27B-DFlash-GGUF to start chatting

Docker Model Runner
How to use Lucebox/Qwen3.6-27B-DFlash-GGUF with Docker Model Runner:
```
docker model run hf.co/Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0
```

Lemonade

How to use Lucebox/Qwen3.6-27B-DFlash-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Lucebox/Qwen3.6-27B-DFlash-GGUF:Q8_0

Run and chat with the model

lemonade run user.Qwen3.6-27B-DFlash-GGUF-Q8_0

List all available models

lemonade list

Qwen3.6-27B DFlash Draft — Q8_0 GGUF

Q8_0 GGUF quantization of the z-lab/Qwen3.6-27B-DFlash draft model, produced for the Lucebox dflash engine (speculative decoding for Qwen3.6-27B-Q4_K_M).

Source: deepsweet/Qwen3.6-27B-DFlash-FP16 (FP16 safetensors mirror of z-lab's BF16)
Tool: dflash/scripts/quantize_draft_q8.py from lucebox-hub
Size: 1.84 GB (53 % of BF16)
Arch: qwen35-dflash-draft, 5 layers, hidden 5120, n_target_layers 5, vocab 248320
Tensors: projection weights → Q8_0, norms → F32 (precision-critical, tiny)
Block size: 16, RoPE θ 1e6, RMS ε 1e-6, MASK token id 248070

Files

File	Size	Purpose
`dflash-draft-3.6-q8_0.gguf`	1.84 GB	The draft model. Pass to dflash via `--draft`

Usage with the Lucebox dflash engine

# 1. Clone + checkout (PR 129 adds Qwen3.6 SWA support)
git clone https://github.com/Luce-Org/lucebox-hub.git
cd lucebox-hub
git fetch origin pull/129/head:pr129 && git checkout pr129
git submodule update --init --recursive

# 2. Build (sm_86+ enables Block-Sparse Attention; sm_75 falls back to ggml flash_attn_ext)
cd dflash
cmake -B build -S . -G Ninja \
  -DCMAKE_BUILD_TYPE=Release \
  -DCMAKE_CUDA_ARCHITECTURES=86 \
  -DDFLASH27B_ENABLE_BSA=ON \
  -DDFLASH27B_TESTS=ON
cmake --build build --target test_dflash -j

# 3. Get the target (Q4_K_M GGUF) and this draft
mkdir -p models/target models/draft
hf download unsloth/Qwen3.6-27B-GGUF --include "*Q4_K_M*.gguf" --local-dir models/target
hf download Lucebox/Qwen3.6-27B-DFlash-GGUF --include "dflash-draft-3.6-q8_0.gguf" --local-dir models/draft

# 4. Run
export DFLASH_TARGET=models/target/Qwen3.6-27B-Q4_K_M.gguf
export DFLASH_DRAFT=models/draft/dflash-draft-3.6-q8_0.gguf
echo "Write a haiku about GPUs." | python3 scripts/run.py --max-ctx 2048 --n-gen 256

The binary autodetects .gguf vs .safetensors from the draft path.

Compatibility

Target: any Qwen3.6-27B-Q4_K_M.gguf (e.g. unsloth/Qwen3.6-27B-GGUF)
The DFlash arch (5 layers + dflash.fc.weight + dflash.hidden_norm.weight) is loaded by gguf_draft_loader.cpp. Quantizing this draft requires the matching converter in dflash/scripts/quantize_draft_q8.py; do not re-quantize with stock llama-quantize — that won't preserve the dflash-specific tensors.

License & attribution

Apache 2.0, inheriting the upstream z-lab license. Original DFlash work and weights by z-lab; FP16 mirror by deepsweet; Q8_0 quantization + repackaging by Lucebox.

Downloads last month: 1,841

GGUF

Model size

2B params

Architecture

qwen35-dflash-draft

Hardware compatibility

8-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Lucebox/Qwen3.6-27B-DFlash-GGUF

Base model

z-lab/Qwen3.6-27B-DFlash

Quantized

(6)

this model