AX Qwen3.6 27B MLX 4-bit MTP

Parameter count: approximately 27.36B logical target parameters (27B class). 4-bit is the target quantization precision, not a 4B model-size claim. The separately packaged MTP sidecar is not included in the target count.

This is an MLX model for Apple Silicon. It packages the 4-bit MLX quantization of Qwen3.6-27B together with an AX Engine-compatible multi-token-prediction (MTP) sidecar.

This repository does not contain PyTorch or GGUF weights.

Model details

• Base model: Qwen/Qwen3.6-27B
• MLX base: mlx-community/Qwen3.6-27B-4bit
• Format: MLX Safetensors
• Quantization: 4-bit affine, group size 64
• Task: image-text-to-text and conversational generation
• Context length: 262,144 tokens in the model configuration
• MTP sidecar: 15 tensors, draft depth up to 3
• Intended hardware: Apple Silicon

The base MLX conversion was produced with mlx-vlm 0.4.4. The MTP tensors come from the official Qwen checkpoint and were normalized with AX Engine's prepare_qwen36_mtp_sidecar.py packaging flow.

Download

hf download AutomatosX/AX-Qwen3.6-27B-MLX-4bit-MTP \
  --local-dir ./AX-Qwen3.6-27B-MLX-4bit-MTP

The repository is approximately 16.93 GB. Ensure that the destination has enough free space before downloading.

Use as an MLX model

For standard MLX-VLM inference without AX MTP acceleration:

pip install -U mlx-vlm

python -m mlx_vlm.generate \
  --model AutomatosX/AX-Qwen3.6-27B-MLX-4bit-MTP \
  --max-tokens 100 \
  --temperature 0.0 \
  --prompt "Describe this image." \
  --image <path-to-image>

The three model-*.safetensors shards are the MLX base weights. Standard MLX-VLM loading uses those shards through model.safetensors.index.json and does not use the AX-specific mtp.safetensors sidecar.

Use with AX Engine MTP

Install AX Engine, download the repository, and serve the local directory:

hf download AutomatosX/AX-Qwen3.6-27B-MLX-4bit-MTP \
  --local-dir ./AX-Qwen3.6-27B-MLX-4bit-MTP

ax-engine serve ./AX-Qwen3.6-27B-MLX-4bit-MTP --port 31418

The MTP package includes:

• mtp.safetensors: the normalized MTP head.
• mtplx_runtime.json: draft-depth and sampler guidance.
• ax_mtp_sidecar_manifest.json: source hashes and transformation provenance.
• config.json: the MLX base configuration with MTP sidecar fields.

Validation and provenance

The sidecar preparation gate recorded:

• Runtime architecture identifier: qwen3-next-mtp
• MTP tensor count: 15
• Norm transform: add 1 to MTP RMSNorm delta weights
• Exactness baseline: maximum absolute difference 0.0 at context length 2048
• Validation platform: macOS on arm64
• Validation timestamp: 2026-06-23

The provenance manifest uses repository-relative paths and public Hugging Face URIs. It intentionally excludes local filesystem paths.

License

This repository follows the upstream Apache License 2.0. See LICENSE and the original Qwen model card for the model's terms, limitations, and responsible-use guidance.