Instructions to use mlx-community/Qwen3.8-Flash-Next-4bit-mtp with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlx-community/Qwen3.8-Flash-Next-4bit-mtp 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("mlx-community/Qwen3.8-Flash-Next-4bit-mtp")
config = load_config("mlx-community/Qwen3.8-Flash-Next-4bit-mtp")

# 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 mlx-community/Qwen3.8-Flash-Next-4bit-mtp with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/Qwen3.8-Flash-Next-4bit-mtp"

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": "mlx-community/Qwen3.8-Flash-Next-4bit-mtp"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent

How to use mlx-community/Qwen3.8-Flash-Next-4bit-mtp 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 "mlx-community/Qwen3.8-Flash-Next-4bit-mtp"

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 mlx-community/Qwen3.8-Flash-Next-4bit-mtp

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use mlx-community/Qwen3.8-Flash-Next-4bit-mtp with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/Qwen3.8-Flash-Next-4bit-mtp"

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 "mlx-community/Qwen3.8-Flash-Next-4bit-mtp" \
  --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"

Qwen3.8-Flash-Next-4bit-mtp (MLX)

mlx-community/Qwen3.8-Flash-Next-4bit plus the model's own multi-token-prediction (MTP) head from Qwen/Qwen3.8-Flash-Next, so it can draft for itself (self-speculative decoding). The 4-bit conversion leaves out the MTP layer; this repository adds it back.

What is in it

The 22 shards of mlx-community/Qwen3.8-Flash-Next-4bit, unchanged (group size 32, 4-bit affine; routers and shared-expert gates at 8-bit, group 64).
One extra shard, model-mtp-00001-of-00001.safetensors (54 tensors, 1.61 GiB): the mtp.* head, fetched from Qwen/Qwen3.8-Flash-Next. The routed-expert gate_up_proj is split into gate and up halves to match the base layout, and the routed experts and the hyper-connection mixing weights are 4-bit affine, group 32; the other head tensors stay bf16.
config.json declares the head (mtp_num_hidden_layers: 1), and model.safetensors.index.json lists the new shard.
mtp_graft_manifest.json records every head tensor's upstream shard, dtype, shape, byte count and sha256.

Downloads last month: 2,256

Safetensors

Model size

180B params

Tensor type

U32

BF16

I64

MLX

Hardware compatibility

4-bit

Model tree for mlx-community/Qwen3.8-Flash-Next-4bit-mtp

Base model

Qwen/Qwen3.8-Flash-Next

Quantized

(628)

this model