Instructions to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-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 ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S
# Run inference directly in the terminal:
llama cli -hf ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S
# Run inference directly in the terminal:
llama cli -hf ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

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 ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S
# Run inference directly in the terminal:
./llama-cli -hf ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

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 ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

Use Docker

docker model run hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

LM Studio
Jan

vLLM

How to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-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": "ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

Ollama
How to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF with Ollama:
```
ollama run hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S
```
Unsloth Desktop

How to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

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": "ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF with Docker Model Runner:
```
docker model run hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S
```

Lemonade

How to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

Run and chat with the model

lemonade run user.Qwen3.8-27B-GSQ-RCO-GGUF-IQ2_S

List all available models

lemonade list

Hermes Agent

How to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-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 ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

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 ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S

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 "ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ2_S" \
  --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"

Qwen3.8-27B · GSQ-RCO GGUFs

Non-uniform GGUF quantizations produced with GSQ and RCO, with a vision projector for multimodal use.

Overview

This repository provides GGUF quantizations of Qwen3.8-27B at three sizes, together with the model's vision projector (mmproj) for multimodal use. In contrast to uniform quantization, which applies a single quantization type to all weight tensors, each model here assigns a separate quantization type to every tensor. The assignment is obtained by a gradient-based search that allocates precision according to per-tensor sensitivity, subject to a total size budget. The resulting files are standard GGUF and run unmodified in llama.cpp, Ollama, and LM Studio.

Method summary. GSQ provides accurate low-bit scalar quantization of each tensor at a given quantization type; RCO assigns the per-tensor quantization types under a size budget. Together they yield a non-uniform GGUF at the requested size.

Method	Description
GSQ (Gumbel-Softmax Quantization, paper, code)	Post-training scalar quantization that jointly learns the per-coordinate grid assignments and the per-group scales via a Gumbel-Softmax relaxation. GSQ closes most of the gap between scalar and vector quantization at 2 to 3 bits while remaining deployable in standard scalar formats such as GGUF.
RCO (Riemannian Constrained Optimization, paper, code)	Assigns one of K quantization types to each of N tensors under a total size budget. The budget constraint is reformulated as a smooth Riemannian manifold in logit space, which permits gradient-based optimization directly on the task loss while enforcing the budget exactly, without constraint-specific hyperparameter tuning.

Both methods were developed at the Deep Algorithms and Systems Lab (DASLab), Institute of Science and Technology Austria.

Available files

Files follow the convention <model>-GSQ-RCO-<type>.gguf, where the suffix names the quantization class; the table lists each file's true whole-file average bit-width. The mmproj file carries the vision encoder and projector at BF16; one copy serves all quantizations.

File	bpw	Size	Notes
`Qwen3.8-27B-GSQ-RCO-IQ2_XS.gguf`	2.50	8.4 GB	Smallest; zero-shot above the BF16 baseline
`Qwen3.8-27B-GSQ-RCO-IQ2_S.gguf`	2.75	9.3 GB	Matches the base model on AIME25
`Qwen3.8-27B-GSQ-RCO-IQ3_XXS.gguf`	3.00	10.1 GB	Recommended; highest quality in this set
`mmproj-Qwen3.8-27B-BF16.gguf`	16	0.9 GB	Vision encoder + projector, for multimodal use

The IQ3_XXS model is the task-lossless operating point: it matches the base model on AIME25 and stays within about one point of it on GPQA-Diamond and LiveCodeBench v6, at one fifth of the BF16 size.

Results

All models are evaluated against the BF16 base model and the Unsloth Dynamic (UD) quantizations of the same base model. We report perplexity on wikitext2, C4, and FineWeb-Edu, the average over five zero-shot tasks (arc_easy, arc_challenge, hellaswag, winogrande, piqa), recovery (zero-shot average relative to BF16), and three reasoning and generation benchmarks: AIME25, GPQA-Diamond, and LiveCodeBench v6. Sizes are those of the files as evaluated.

Variant	bpw	GB	wiki↓	c4↓	fw↓	ZS avg↑	recovery	AIME25↑	GPQA-D↑	LCB v6↑
BF16	16.00	53.8	7.05	11.45	8.14	74.34	100.0%	100.00	89.90	85.71
GSQ-RCO IQ2_XS	2.50	8.4	7.69	12.98	9.19	74.54	100.3%	96.67	84.85	76.57
GSQ-RCO IQ2_S	2.75	9.3	7.39	12.40	8.80	75.70	101.8%	100.00	86.36	82.29
GSQ-RCO IQ3_XXS	3.00	10.1	7.20	12.13	8.59	74.81	100.6%	100.00	88.89	84.57
UD-IQ2_S	2.49	8.4	8.02	12.78	9.08	73.80	99.3%	86.67	76.26	72.00
UD-Q2_K_XL	2.88	9.8	7.54	12.25	8.69	74.37	100.0%	100.00	86.87	82.28

At 3.0 bpw (IQ3_XXS), the model matches the base on AIME25 (100.00) and stays within about one point of it on GPQA-Diamond (88.89 vs 89.90) and LiveCodeBench v6 (84.57 vs 85.71), at 10.1 GB. IQ2_S restores AIME25 to 100.00 at 9.3 GB and exceeds the BF16 zero-shot average (75.70 vs 74.34); against UD-Q2_K_XL it is 0.5 GB smaller at near-identical task scores. At matched file size (8.4 GB), IQ2_XS leads UD-IQ2_S by 10.00 points on AIME25, 8.59 on GPQA-Diamond, and 4.57 on LiveCodeBench v6.

Usage

llama.cpp

# download (requires: pip install -U "huggingface_hub[cli]")
hf download ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF Qwen3.8-27B-GSQ-RCO-IQ3_XXS.gguf --local-dir .

llama-cli -m Qwen3.8-27B-GSQ-RCO-IQ3_XXS.gguf -p "Explain mixed-precision quantization." -ngl 99

Vision (multimodal)

hf download ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF mmproj-Qwen3.8-27B-BF16.gguf --local-dir .

llama-mtmd-cli -m Qwen3.8-27B-GSQ-RCO-IQ3_XXS.gguf \
  --mmproj mmproj-Qwen3.8-27B-BF16.gguf \
  --image photo.jpg -p "Describe this image."

The projector was converted directly from the base checkpoint and verified against these quantizations.

Ollama

ollama run hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF   # pick the file matching your memory budget

LM Studio

Search the repo name, then pick a GSQ-RCO-* build from the file list.

Quantization procedure

Per-tensor database. Each weight tensor is quantized at every candidate GGUF quantization type with GSQ, yielding a searchable database of quantized tensor variants.
RCO search. The budget-constrained Riemannian search assigns one quantization type per tensor such that the whole-file average bit-width meets the target.
Assembly. The selected per-tensor variants are stitched into a single standard GGUF file.

Reference implementations: GSQ at IST-DASLab/GSQ and RCO at IST-DASLab/RCO.

Citation

If you use these models or methods, please cite both papers:

@article{gsq2026,
  title  = {GSQ: Highly-Accurate Low-Precision Scalar Quantization for LLMs via Gumbel-Softmax Sampling},
  author = {Dadgarnia, Alireza and Tabesh, Soroush and Nikdan, Mahdi and Helcig, Michael and Kurtic, Eldar and Kleinegger, Maximilian and Alistarh, Dan},
  journal= {arXiv preprint arXiv:2604.18556},
  year   = {2026}
}
@article{rco2026,
  title  = {Model Compression with Exact Budget Constraints via Riemannian Manifolds},
  author = {Helcig, Michael and Alistarh, Dan},
  journal= {arXiv preprint arXiv:2605.00649},
  year   = {2026}
}

License

These quantized weights inherit the license of the base model (Qwen3.8-27B). The GSQ-RCO tooling is released by the Deep Algorithms and Systems Lab under its repository license.

_{Built with GSQ and RCO at the Deep Algorithms and Systems Lab · Institute of Science and Technology Austria}