Instructions to use WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit 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("WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit")

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 WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit"

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": "WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit 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 "WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit"

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 WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit

Run Hermes

hermes

OpenClaw new

How to use WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit"

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 "WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit" \
  --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 WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Nanbeige4.2-3B Heretic MLX DWQ 4-bit

A 4-bit, group-size-32 Distilled Weight Quantization (DWQ) release of WaveCut/Nanbeige4.2-3B-heretic.

This repository includes a small trusted-code MLX-LM adapter because Nanbeige reuses 22 physical decoder layers over two loops and is not a standard Llama layout at runtime. The adapter preserves shared weights while allocating 44 independent KV caches, one for each loop/layer execution. Unsupported optional Nanbeige architectures are rejected explicitly.

DWQ calibration

4 bits, group size 32.
1,024 training samples and 32 validation samples.
Maximum sequence length: 1,025 tokens.
Seed: 20260722.
Corpus: 528 agentic trajectories plus 528 coding-reasoning examples, deterministically shuffled.
Corpus SHA-256: a7cfdbe02c124304bf1282bbd5ed7162bfa72dec6750b60ed2d3a68000c7a554.
Agentic source: TIGER-Lab/SWE-QA-Pro-SFT-Trajectories at b8f5b8a8dcf90bca8b6d70adedac0d20dca02b86.
Coding source: nvidia/OpenCodeReasoning at 20a1ca19c0d050fe9057fc08339d6b370ec1c67a.

Validation loss	Value
Initial RTN	0.284
Final DWQ	0.043

MLX-LM revision: cf10f962b7a20e63a6df43dbf0faf06070153d40.

Usage

The model file is repository code, so load it only after reviewing nanbeige_mlx.py and pass --trust-remote-code.

mlx_lm.generate \
  --model WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit \
  --trust-remote-code \
  --prompt "Implement a bounded async worker pool in Python." \
  --max-tokens 256

Exact artifact hashes and clean-load smoke-test results are recorded in release-manifest.json.

Downloads last month: 480

Safetensors

Model size

0.8B params

Tensor type

BF16

U32

MLX

Hardware compatibility

4-bit

Model tree for WaveCut/Nanbeige4.2-3B-heretic-MLX-DWQ-4bit

Base model

Nanbeige/Nanbeige4.2-3B-Base

Finetuned

Nanbeige/Nanbeige4.2-3B

Finetuned

WaveCut/Nanbeige4.2-3B-heretic

Quantized

(4)

this model