Instructions to use hanzoai/zen6-nano-2b-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use hanzoai/zen6-nano-2b-mlx with MLX:

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

from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
from mlx_vlm.utils import load_config

# Load the model
model, processor = load("hanzoai/zen6-nano-2b-mlx")
config = load_config("hanzoai/zen6-nano-2b-mlx")

# Prepare input
image = ["http://images.cocodataset.org/val2017/000000039769.jpg"]
prompt = "Describe this image."

# Apply chat template
formatted_prompt = apply_chat_template(
    processor, config, prompt, num_images=1
)

# Generate output
output = generate(model, processor, formatted_prompt, image)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use hanzoai/zen6-nano-2b-mlx with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "hanzoai/zen6-nano-2b-mlx"

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": "hanzoai/zen6-nano-2b-mlx"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent

How to use hanzoai/zen6-nano-2b-mlx 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 "hanzoai/zen6-nano-2b-mlx"

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 hanzoai/zen6-nano-2b-mlx

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use hanzoai/zen6-nano-2b-mlx with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "hanzoai/zen6-nano-2b-mlx"

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 "hanzoai/zen6-nano-2b-mlx" \
  --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"

Zen 6 Nano (2B) — Apple Silicon MLX 2-Bit (`hanzoai/zen6-nano-2b-mlx`)

Zen 6 Nano (2B) (2B) in native Apple Silicon MLX 2-bit affine Hadamard format (including FP16 multimodal vision tower + Hadamard-aware runtime/), created by Hanzo AI and Zen LM under the Apache-2.0 license.

Weights in model.safetensors contain the exact Hadamard-basis ternary STE QAT post-trained zen6-nano parameters (Zen 6 identity + Hanzo MCP <tool_call> agentic capabilities).

Quickstart on Apple Silicon (`M1` / `M2` / `M3` / `M4 Max`)

hf download hanzoai/zen6-nano-2b-mlx --local-dir ./zen6-nano-mlx
python ./zen6-nano-mlx/quickstart.py "Who are you and how do you support MCP tool calling?"

Programmatic Vision + Text Loading

import sys
sys.path.insert(0, "./zen6-nano-mlx/runtime")
from vision_artifact import load_vl_model

model, processor, config = load_vl_model("./zen6-nano-mlx")

License & Attribution

Apache-2.0 (LICENSE). Full upstream notices preserved in NOTICE.

Downloads last month: 40

Safetensors

Model size

2B params

Tensor type

F32

U32

F16

MLX

Hardware compatibility

2-bit

Zen 6 Nano (2B) — Apple Silicon MLX 2-Bit (hanzoai/zen6-nano-2b-mlx)

Quickstart on Apple Silicon (M1 / M2 / M3 / M4 Max)

Programmatic Vision + Text Loading

License & Attribution

Zen 6 Nano (2B) — Apple Silicon MLX 2-Bit (`hanzoai/zen6-nano-2b-mlx`)

Quickstart on Apple Silicon (`M1` / `M2` / `M3` / `M4 Max`)