Instructions to use mlx-community/GLM-5.2-mxfp4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlx-community/GLM-5.2-mxfp4 with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/GLM-5.2-mxfp4")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use mlx-community/GLM-5.2-mxfp4 with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/GLM-5.2-mxfp4"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "mlx-community/GLM-5.2-mxfp4"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use mlx-community/GLM-5.2-mxfp4 with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/GLM-5.2-mxfp4"

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 mlx-community/GLM-5.2-mxfp4

Run Hermes

hermes

OpenClaw new

How to use mlx-community/GLM-5.2-mxfp4 with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/GLM-5.2-mxfp4"

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 "mlx-community/GLM-5.2-mxfp4" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

MLX LM

How to use mlx-community/GLM-5.2-mxfp4 with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "mlx-community/GLM-5.2-mxfp4"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "mlx-community/GLM-5.2-mxfp4"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "mlx-community/GLM-5.2-mxfp4",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Successful 512GB Mac Studio run with Pedro PR branch

by hugfaceme - opened 16 days ago

Discussion

hugfaceme

16 days ago

I got mlx-community/GLM-5.2-mxfp4 running locally on a 512GB Mac Studio.

My original failure path was the same class of issue: missing self_attn.indexer.* / missing indexer parameters. In my case the model download was fine , the key was using the patched mlx-lm runtime for GLM-5.2 DSA/indexer sharing.

Working setup:

512GB Mac Studio
mlx-community/GLM-5.2-mxfp4
clean isolated venv
MLX / mlx-metal 0.31.2
Pedro’s mlx-lm branch: glm-moe-dsa-indexer-sharing
jinja2
sudo sysctl iogpu.wired_limit_mb=480000
tested first with mlx_lm.generate
then ran local mlx_lm.server
Open WebUI connected to http://127.0.0.1:8080/v1
One small practical gotcha: I needed jinja2. Without it, the tokenizer chat template failed before generation.

Here is exact practical steps here:

https://gist.github.com/LocalAiCherry/6c769b529ebaebc6088449f63025676d

hugfaceme

16 days ago

One small practical dependency I had to add was jinja2, a Python template library used by the tokenizer chat template. Without it, generation stopped before the model could answer.

thibautrey

15 days ago

what speed to you get ? Do you see much difference with the official unquantized model ?

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