Instructions to use stamsam/Qwen3.6-28B-REAP-oQ4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use stamsam/Qwen3.6-28B-REAP-oQ4 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("stamsam/Qwen3.6-28B-REAP-oQ4")

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 stamsam/Qwen3.6-28B-REAP-oQ4 with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "stamsam/Qwen3.6-28B-REAP-oQ4"

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": "stamsam/Qwen3.6-28B-REAP-oQ4"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use stamsam/Qwen3.6-28B-REAP-oQ4 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 "stamsam/Qwen3.6-28B-REAP-oQ4"

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 stamsam/Qwen3.6-28B-REAP-oQ4

Run Hermes

hermes

MLX LM

How to use stamsam/Qwen3.6-28B-REAP-oQ4 with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "stamsam/Qwen3.6-28B-REAP-oQ4"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "stamsam/Qwen3.6-28B-REAP-oQ4"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "stamsam/Qwen3.6-28B-REAP-oQ4",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Does MLX support smaller quantizations?

by robbintt - opened 19 days ago

Discussion

robbintt

19 days ago

•

edited 19 days ago

Does MLX support smaller quantizations?

I am using https://huggingface.co/mradermacher/Qwen3.6-28B-REAP-i1-GGUF?show_file_info=Qwen3.6-28B-REAP.i1-IQ3_XXS.gguf which is 11.2 GB on a 16GB mac mini. It's working great but I would like to check out a MLX quant if possible.

stamsam

Owner 17 days ago

Yes, it does. I’m working on JANG compression style tweaks for this model. I tested it on a Gemma REAP ( https://huggingface.co/stamsam/Gemma-4-21B-REAP-JANG3M-MLX ) to make them smaller. Take a peek at it. Let me know if it works on your Mac while I work on this one.

robbintt

17 days ago

Thank you, that will definitely fit. I will test it while I wait for that one. I think about 11.3 GB is solid, even smaller is better so I can still use browser and terminal. But, its hard to know what's possible and what the quality tradeoffs are.

stamsam

Owner 17 days ago

The quality in the one I recommend you try Gemma real above out should be way better than the Q1 you’re ruining, and I’ll let you know when I finish the Qwen jang reap. Do you use X ? If so, I’ll message you there. My profile is @stamatiou justdm

robbintt

17 days ago

Thanks, followed you there, got your followback.

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