Instructions to use EigenLabs/Qwen3.8-27B-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use EigenLabs/Qwen3.8-27B-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("EigenLabs/Qwen3.8-27B-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 EigenLabs/Qwen3.8-27B-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 "EigenLabs/Qwen3.8-27B-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": "EigenLabs/Qwen3.8-27B-4bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

OpenClaw new

How to use EigenLabs/Qwen3.8-27B-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 "EigenLabs/Qwen3.8-27B-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 "EigenLabs/Qwen3.8-27B-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 EigenLabs/Qwen3.8-27B-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 "EigenLabs/Qwen3.8-27B-4bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "EigenLabs/Qwen3.8-27B-4bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "EigenLabs/Qwen3.8-27B-4bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use EigenLabs/Qwen3.8-27B-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 "EigenLabs/Qwen3.8-27B-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 EigenLabs/Qwen3.8-27B-4bit

Run Hermes

hermes

Atomic Chat

Qwen3.8-27B-4bit

This is a 4-bit MLX version of Qwen/Qwen3.8-27B.

We made it from revision 1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0 of the official model. We used mlx_lm.convert with mlx 0.32.0. The quantization is affine, 4 bits, group size 64.

Architecture

The model has 64 layers.
Each group of 4 layers has 3 linear-attention layers and 1 full-attention layer.
The hidden size is 5120.
The model has 24 attention heads and 4 KV heads. The head dimension is 256.
The vocabulary has 248,320 tokens.
The model type is qwen3_5.

MTP head

The official model has an MTP (multi-token-prediction) head. The head has 15 tensors. This repository does not contain these tensors. They are in EigenLabs/Qwen3.8-27B-MTP-bf16. That repository keeps them in bfloat16.

Role

The Qwen 3.8 MTP challenge uses this model as its reference. The track is qwen3.8-27b-mtp-v1. The challenge repository is Layr-Labs/qwen-3.8-mtp-challenge. The challenge points to one fixed revision of this repository. The challenge compares its files against this revision, byte for byte.

Usage

pip install mlx-lm
mlx_lm.generate --model EigenLabs/Qwen3.8-27B-4bit --prompt "Hello"

Downloads last month: 49

Safetensors

Model size

4B params

Tensor type

BF16

U32

MLX

Hardware compatibility

4-bit

Model tree for EigenLabs/Qwen3.8-27B-4bit

Base model

Qwen/Qwen3.8-27B

Quantized

(401)

this model