Instructions to use Beinsezii/GLM-4.6V-GGUF-HALO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Beinsezii/GLM-4.6V-GGUF-HALO 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 Beinsezii/GLM-4.6V-GGUF-HALO:F16
# Run inference directly in the terminal:
llama cli -hf Beinsezii/GLM-4.6V-GGUF-HALO:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Beinsezii/GLM-4.6V-GGUF-HALO:F16
# Run inference directly in the terminal:
llama cli -hf Beinsezii/GLM-4.6V-GGUF-HALO: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 Beinsezii/GLM-4.6V-GGUF-HALO:F16
# Run inference directly in the terminal:
./llama-cli -hf Beinsezii/GLM-4.6V-GGUF-HALO: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 Beinsezii/GLM-4.6V-GGUF-HALO:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Beinsezii/GLM-4.6V-GGUF-HALO:F16

Use Docker

docker model run hf.co/Beinsezii/GLM-4.6V-GGUF-HALO:F16

LM Studio
Jan
Ollama
How to use Beinsezii/GLM-4.6V-GGUF-HALO with Ollama:
```
ollama run hf.co/Beinsezii/GLM-4.6V-GGUF-HALO:F16
```
Unsloth Desktop

How to use Beinsezii/GLM-4.6V-GGUF-HALO with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Beinsezii/GLM-4.6V-GGUF-HALO: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": "Beinsezii/GLM-4.6V-GGUF-HALO:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Beinsezii/GLM-4.6V-GGUF-HALO with Docker Model Runner:
```
docker model run hf.co/Beinsezii/GLM-4.6V-GGUF-HALO:F16
```

Lemonade

How to use Beinsezii/GLM-4.6V-GGUF-HALO with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Beinsezii/GLM-4.6V-GGUF-HALO:F16

Run and chat with the model

lemonade run user.GLM-4.6V-GGUF-HALO-F16

List all available models

lemonade list

Hermes Agent

How to use Beinsezii/GLM-4.6V-GGUF-HALO with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Beinsezii/GLM-4.6V-GGUF-HALO: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 Beinsezii/GLM-4.6V-GGUF-HALO:F16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Beinsezii/GLM-4.6V-GGUF-HALO with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Beinsezii/GLM-4.6V-GGUF-HALO: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 "Beinsezii/GLM-4.6V-GGUF-HALO: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"

Quant optimized for quality / speed on a Strix Halo 128GiB system. Possibly also beneficial on DGX Spark and similar systems.

The TL;DR is this quant achieves both superior quality and speed to homogenous Q6_K

Q6_K

model	size	params	backend	ngl	n_ubatch	fa	test	t/s
glm4moe ?B Q6_K	89.58 GiB	106.85 B	ROCm	999	1024	1	pp2048	187.87 ± 0.00
glm4moe ?B Q6_K	89.58 GiB	106.85 B	ROCm	999	1024	1	tg256	16.73 ± 0.00
glm4moe ?B Q6_K	89.58 GiB	106.85 B	ROCm	999	1024	1	pp2048 @ d8192	120.83 ± 0.00
glm4moe ?B Q6_K	89.58 GiB	106.85 B	ROCm	999	1024	1	tg256 @ d8192	13.41 ± 0.00

This quant

model	size	params	backend	ngl	n_ubatch	fa	test	t/s
glm4moe ?B Q8_0	90.80 GiB	106.85 B	ROCm	999	1024	1	pp2048	296.28 ± 0.00
glm4moe ?B Q8_0	90.80 GiB	106.85 B	ROCm	999	1024	1	tg256	15.58 ± 0.00
glm4moe ?B Q8_0	90.80 GiB	106.85 B	ROCm	999	1024	1	pp2048 @ d8192	160.92 ± 0.00
glm4moe ?B Q8_0	90.80 GiB	106.85 B	ROCm	999	1024	1	tg256 @ d8192	12.69 ± 0.00

What this quant does is move some hot layers (attention, shared expert) to q8_0 for faster processing. Basically Q6_K is the optimal size for the Halo but it's also the slowest quant, made worse by the fact that it performs poorly on MMQ kernels which GLM4Moe always uses due to its high exp count. For detailed RDNA 3.0 benchmarks you can view my kernel selection PR here as well Johnathan's follow-up RDNA3.5 version here

Additionally the context should still fit ≥90k with room for a graphical desktop assuming a large TTM was set. Completely headless you might be able to reach full 128k.

Everything above assumes you're running ROCm, not Vulkan. Vulkan being faster is a myth. While it might look like +15% for tg512, when run at even a modest context depth, the speed becomes catastrophic

This quant, Vulkan

model	size	params	backend	ngl	n_ubatch	fa	test	t/s
glm4moe ?B Q8_0	90.80 GiB	106.85 B	Vulkan	999	1024	1	pp2048	244.54 ± 0.00
glm4moe ?B Q8_0	90.80 GiB	106.85 B	Vulkan	999	1024	1	tg256	17.18 ± 0.00
glm4moe ?B Q8_0	90.80 GiB	106.85 B	Vulkan	999	1024	1	pp2048 @ d8192	33.08 ± 0.00
glm4moe ?B Q8_0	90.80 GiB	106.85 B	Vulkan	999	1024	1	tg256 @ d8192	13.71 ± 0.00