Instructions to use agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-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 agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-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 agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16
# Run inference directly in the terminal:
llama cli -hf agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16
# Run inference directly in the terminal:
llama cli -hf agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

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 agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16
# Run inference directly in the terminal:
./llama-cli -hf agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

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 agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

Use Docker

docker model run hf.co/agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

LM Studio
Jan

vLLM

How to use agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-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": "agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

Ollama
How to use agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF with Ollama:
```
ollama run hf.co/agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16
```
Unsloth Desktop

How to use agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

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": "agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF with Docker Model Runner:
```
docker model run hf.co/agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16
```

Lemonade

How to use agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

Run and chat with the model

lemonade run user.Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF-F16

List all available models

lemonade list

Hermes Agent

How to use agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-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 agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

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 agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16

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 "agentionai/Qwen3.8-Flash-Next-ROCmFP4-FAST-imatrix-GGUF:F16" \
  --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"

Superb Work

by nosell4211 - opened 4 days ago

Discussion

nosell4211

4 days ago

Thank you for this, it runs beautifully on my Strix Halo, with 262000 of context, on fedora 44. A huge improvement in speed over the original and with great results.

Prefill at depth is the standout. 452 t/s at 2k decaying to 234 t/s at 68k — roughly −3.3 t/s per 1k, and still faster at 68k than the Unsloth quant managed at 8.7k.
Slot selection by LCP similarity hit 0.957, 0.999, 0.991, 0.995, 0.985, 0.993 across the session; task 38 re-prefilled only 3,136 tokens against a ~70k context.
Across seven tasks all sitting between 77k and 91k depth, decode ranged 14.90 to 32.12 t/s

jkimber

about 7 hours ago

Tested Qwen3.8-Flash-Next-ROCmFP4-FAST-v2-ple16.gguf with a GMKTEC Evo-X2 (Ryzen AI MAX+ 395, 128 GB LPDDR5X) running Fedora 44 (kernel 6.14.whatever numbers make it the most current version as of September 2, 2026) using llama.cpp over Vulkan (RADV), and the results are incredible!

Running entirely within unified iGPU memory required two key adjustments for the Linux driver. Maybe very specific to my machine and operating system...

amdgpu.gttsize=126976 ttm.pages_limit=28311552 via GRUB to unlock >100GB addressable GTT space

GGML_VK_FORCE_MAX_ALLOCATION_SIZE=2147483648 (2GB chunks) at launch to eliminate radv: Not enough memory for command submission / ErrorDeviceLost panics

If there's a better way, I'm all ears. It's the only solution I got working.

Generation sits rock-solid at between 26 and 28 t/s, which matches the published benchmarks. I'm getting prefill speeds around 58 t/s.
Full multimodal support (--mmproj) works out of the box: OCR, document structure parsing, and visual reasoning, all great.

The depth of domain knowledge and physical reasoning on technical camera optics rivals frontier models. Writes beautiful Typst formats that allows my fine tuned and harnessed IBM Granite model to tweak to match my template structures quickly, translates Japanese to English and back with minimal need to edit (aside from personal taste) -- can't wait to continue building on this!

Fantastic work on these kernels and quantization tables! Looking forward to see more Strix Halo love from you!

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment