Instructions to use Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS 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 Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS
# Run inference directly in the terminal:
llama cli -hf Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS
# Run inference directly in the terminal:
llama cli -hf Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

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 Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS
# Run inference directly in the terminal:
./llama-cli -hf Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

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 Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

Use Docker

docker model run hf.co/Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

LM Studio
Jan
Ollama
How to use Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS with Ollama:
```
ollama run hf.co/Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS
```
Unsloth Desktop

How to use Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

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": "Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS with Docker Model Runner:
```
docker model run hf.co/Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS
```

Lemonade

How to use Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

Run and chat with the model

lemonade run user.Qwen3.8-Flash-Next-R9V-IQ4_XS-UD-IQ4_XS

List all available models

lemonade list

Hermes Agent

How to use Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

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 Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS

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 "Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS:UD-IQ4_XS" \
  --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 Flash Next R9V IQ4_XS

Ready-to-arrange model bundle for the R9V dual-RDNA4 inference profile.

Contents and provenance

target/: Unsloth UD-IQ4_XS target GGUF shards from unsloth/Qwen3.8-Flash-Next-GGUF.
vision/: ggml-org Q8_0 vision projector. This projector was not quantized by R9V.
mtp/: R9V-assembled minimal MTP checkpoint. Dense/nonexpert tensors come from the official BF16 checkpoint; routed experts come from the official block-FP8 checkpoint. R9V did not train these weights.
metadata/: official Qwen tokenizer, processor, and model configuration.
manifests/: the reference dual-R9700 hot-expert placement.
sources.lock.json: exact upstream revisions, sizes, and hashes.

Unsloth and ggml-org are credited for the target quantization and vision projector. Qwen remains the model author and upstream rights holder.

PLE table

The 26.82 GiB per_layer_token_embd.weight payload is already present inside target shard 2 and is intentionally not uploaded again. Extract it to the fast SSD with R9V's metadata-driven tool. The extraction utility is shipped in the R9V source repository, not this model repository; follow the complete build, extraction, and launch instructions.

Reference configuration

Two Radeon R9700 32 GiB GPUs, TP2.
128 GiB DDR5.
MTP depth 2.
131,072-token capacity, BF16 QSA KV.
Q8 vision input, one image/request.
SSD PLE and tiered expert placement.

The immutable 22-file model package is public and remotely hash-verified. The R9V runtime has passed its reference-machine performance qualification and remains a release candidate until the documented package installation passes from a clean host. See the R9V qualification report for exact benchmark cells and protocol.

License

These model artifacts are distributed under Qwen Community License 1.0; see LICENSE. The R9V Apache-2.0 code license does not apply to model weights. Users are responsible for reviewing the Qwen license, including its separate terms for certain commercial MaaS/AI-work-assistant uses and scale thresholds. Artifact attribution and exact upstream revisions are recorded in THIRD_PARTY_NOTICES.md.

Downloads last month: 816

GGUF

Model size

177B params

Architecture

qwen4exp

Hardware compatibility

4-bit

Inference Providers NEW

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

Model tree for Dyluhn/Qwen3.8-Flash-Next-R9V-IQ4_XS

Base model

Qwen/Qwen3.8-Flash-Next

Quantized

(252)

this model