Instructions to use JANGQ-AI/MiniMax-M3-REAP40-JANG_2L with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use JANGQ-AI/MiniMax-M3-REAP40-JANG_2L 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("JANGQ-AI/MiniMax-M3-REAP40-JANG_2L")

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 JANGQ-AI/MiniMax-M3-REAP40-JANG_2L with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "JANGQ-AI/MiniMax-M3-REAP40-JANG_2L"

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": "JANGQ-AI/MiniMax-M3-REAP40-JANG_2L"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use JANGQ-AI/MiniMax-M3-REAP40-JANG_2L 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 "JANGQ-AI/MiniMax-M3-REAP40-JANG_2L"

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 JANGQ-AI/MiniMax-M3-REAP40-JANG_2L

Run Hermes

hermes

OpenClaw new

How to use JANGQ-AI/MiniMax-M3-REAP40-JANG_2L with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "JANGQ-AI/MiniMax-M3-REAP40-JANG_2L"

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 "JANGQ-AI/MiniMax-M3-REAP40-JANG_2L" \
  --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 JANGQ-AI/MiniMax-M3-REAP40-JANG_2L with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "JANGQ-AI/MiniMax-M3-REAP40-JANG_2L"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "JANGQ-AI/MiniMax-M3-REAP40-JANG_2L"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "JANGQ-AI/MiniMax-M3-REAP40-JANG_2L",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Grate work , thank you

by infinityai - opened 22 days ago

Discussion

infinityai

22 days ago

•

edited 22 days ago

Grate work , thank you

Was just looking for the reap quants

How did you do the REAP?

Are you planning to make a REAP verson of GLM 5.2 ?and on Kimi code 2.7 ?

It would be amazing if that can also fit in 128gb macs also

jangq

JANG_Q - GGUF for MLX org 22 days ago

@infinityai - saliency reap using a custom agentic usage calibration set, yes i want to do a glm 5.2 - i see your work with REAPs, if you ever wanna work together lmk!

infinityai

18 days ago

Yes am up for working together, sounds good.

I found this version we could do a test on with your quantisation

https://huggingface.co/pipenetwork/GLM-5.2-REAP50-MLX-4bit

ideally we would use the 16 bit REAP weights if we can get them , but for now we could test the 4bit he has made with the 50 percent reap on GLM 5.2

infinityai

15 days ago

This is a good one to test, this is BF16 with 34 percent REAP , off GLM 5.2

https://huggingface.co/0xSero/GLM-5.2-REAP-504B-GGUF

redhat-ruse

JANG_Q - GGUF for MLX org 15 days ago

How about a jangq1 reap that would enable the model to fit on an 128 or even 64gb macbook pro. I would be willing to test the extremes - I have a 36gb macbook pro and would love to get this model running on it as it would so fun even if I had to use ssd streaming or smelt moe.

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