Instructions to use aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw 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("aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw")

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 aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw"

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": "aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw 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 "aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw"

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 aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw"

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 "aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw" \
  --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"

Qwen3.6-35B-A3B — MLX VQ @3.4 bpw: beats scalar 3.5bpw at a smaller size

Vector-quantized (trained codebooks + GPTQ compensation) MLX build of Qwen/Qwen3.6-35B-A3B — the quality-first sibling of Qwen3.6-35B-A3B-MLX-VQ-2.6bpw. ~15.0 GB, experts uniformly VQ-3bit (d=4, K=4096 Qwen-native codebook).

⚠️ Not loadable by stock mlx-lm / LM Studio / Ollama / oMLX — use the bundled loader + Metal kernel in code/ (pure Python, mlx>=0.31, no build step).

Quality (PPL vs bf16, ja-think / multilingual holdouts)

build	size	think-ja	multilingual
GPTQ-3.5bpw (scalar)	15.35 GB	+6.6%	+6.9%
this (VQ-3.4bpw)	15.0 GB	+6.0%	+5.8%
VQ-2.6bpw	11.53 GB	+12.4%	+12.2%
GPTQ-2.7bpw (scalar)	12.03 GB	+17.5%	+15.4%
VQ-2.4bpw — for 16 GB Macs	10.49 GB	+19.5%	+21.6%

Smaller and better than the scalar 3.5bpw build on both yardsticks — the trained codebook's edge, now at the quality end of the curve. (Measured with an RTN spine; the shipped artifact's spine is GPTQ-compensated, i.e. slightly better than the table.)

Recipe

Experts (91.7% of params): uniform VQ 3-bit — d=4 subvectors, K=4096 Qwen-native codebook (k-means on group-normalized subvectors), GPTQ error compensation with pooled per-layer input Hessians (ja-centric calibration). 12-bit packed indices + fp16 group scales (eff 3.125 bpw on experts).
Spine: GPTQ 4bit gs64 (linear-attention, attention, shared experts), lm_head 6bit, routers 8bit, embeddings 4bit. Vision tower: unquantized bf16.

Run (Apple Silicon, ≥24 GB unified memory)

pip install "mlx>=0.31" "mlx-lm>=0.31"
hf download aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw --local-dir qwen-vq35

python qwen-vq35/code/vq_serve.py --model qwen-vq35 --port 8091   # OpenAI-compatible
python qwen-vq35/code/vq_generate.py qwen-vq35 100                # one-shot test

Decode is the all-12-bit-unpack path: measured 52 tok/s on an M-series 48 GB (vs 66 for the 2.6bpw build — more expert bytes per token).

Provenance / license

Base model © Alibaba Cloud, Apache-2.0. Pipeline: loss-aware allocation + GPTQ-VQ encode (OneCompression); serving adapter = vqmoe model #3.

Downloads last month: 115

Safetensors

Model size

6B params

Tensor type

U32

I32

BF16

F16

MLX

Hardware compatibility

4-bit

Model tree for aquaman164/Qwen3.6-35B-A3B-MLX-VQ-3.4bpw

Base model

Qwen/Qwen3.6-35B-A3B

Quantized

(828)

this model