Instructions to use RMDWLLC/Jah-7.0-Mini-MLX-8bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RMDWLLC/Jah-7.0-Mini-MLX-8bit 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("RMDWLLC/Jah-7.0-Mini-MLX-8bit")

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 RMDWLLC/Jah-7.0-Mini-MLX-8bit with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "RMDWLLC/Jah-7.0-Mini-MLX-8bit"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "RMDWLLC/Jah-7.0-Mini-MLX-8bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use RMDWLLC/Jah-7.0-Mini-MLX-8bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "RMDWLLC/Jah-7.0-Mini-MLX-8bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "RMDWLLC/Jah-7.0-Mini-MLX-8bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "RMDWLLC/Jah-7.0-Mini-MLX-8bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use RMDWLLC/Jah-7.0-Mini-MLX-8bit 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 "RMDWLLC/Jah-7.0-Mini-MLX-8bit"

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 RMDWLLC/Jah-7.0-Mini-MLX-8bit

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use RMDWLLC/Jah-7.0-Mini-MLX-8bit with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "RMDWLLC/Jah-7.0-Mini-MLX-8bit"

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 "RMDWLLC/Jah-7.0-Mini-MLX-8bit" \
  --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"

Jah 7.0 Mini (MLX, 8-bit)

Jah 7.0 Mini is the small member of the Jah 7.0 family from RMDW LLC, built to run on your own Mac. Same Jah behavior as the big model: it answers the questions other models refuse, it calls tools, and it can think before it answers. No cloud, no account, nothing leaves your machine.

The full Jah 7.0 needs a data-center class machine. Jah 7.0 Mini runs on a laptop.

At a glance

Size: 27 GB on disk (8-bit). Runs on Macs with 64 GB of memory or more.
Speed: about 24 tokens per second on an Apple M5 Max (4-bit).
Uncensored: refusal behavior removed at the weight level.
Tool calling and a thinking mode (turn it on for hard problems, off for instant answers).
Knows who it is: it introduces itself as Jah, made by RMDW. Your own system prompt still works on top of that.

Quality

On the censorship battery used for every Jah release (historical, political, and mechanical questions that stock models refuse), Jah 7.0 Mini answers 6 of 6. It passes the same capability checks as the big models: exact arithmetic, tool calls, multi-turn memory, and thinking mode.

Run it

pip install mlx-lm
mlx_lm.generate --model RMDWLLC/Jah-7.0-Mini-MLX-8bit --prompt "Who are you?"

Or serve it with an OpenAI-compatible API on your Mac:

mlx_lm.server --model RMDWLLC/Jah-7.0-Mini-MLX-8bit --port 8080

It also loads in LM Studio (MLX engine). Thinking is controlled with chat_template_kwargs: {"enable_thinking": true}.

The Jah family

Jah 7.0: the flagship. 1M-token context, vision, runs on 4x 96 GB GPUs. Serves Kiyomi.
Jah 7.0 Mini (this repo): built from a smaller open base, aligned by RMDW, for your own Mac. 4-bit and 8-bit.

Want Jah without running anything? It powers Kiyomi, private AI set up and run for you.

License

Apache 2.0, the license of the open base this model is built from. See LICENSE.

RMDW LLC, Georgia. rmdw.ai · kiyomibot.ai

Downloads last month: -

Safetensors

Model size

27B params

Tensor type

U32

BF16

MLX

Hardware compatibility

8-bit