AX-Qwen3.6-35B-A3B-MLX-AXQ-4bit-MTP

An AXQuant (AXQ) mixed-precision MLX checkpoint for Apple Silicon, converted directly from the BF16 source model. The language path is quantized while the multi-token-prediction (MTP) head and vision tower are preserved as BF16 sidecars when present.

AXQuant checkpoint Tier 1 certified on df-macbookpro-m5 for Hub commit a549387d5b812c6f6cbdb0ebde37adb3b3f4a2bc. Agent-coding and general quality retention both 1.000 vs mlx-community/Qwen3.6-35B-A3B-4bit. Size vs uniform-4 is 1.132× under a class budget of 1.15. Certificate.

MTP acceleration Tier 2 is not certified. Formal greedy exactness is achievable with MoE MTP loaded, but decode-heavy speed gates (≥1.20× weighted / ≥1.10× prompt-median) are not met (agent ~0.95× / 0.80×; long ~0.91× / 0.93×). Product default remains direct fallback. Certification index.

Model details

This repository contains MLX Safetensors. It does not contain PyTorch or GGUF weights.

Choosing an AXQ pack

AXQ names describe a storage-budget product class, not one uniform precision applied to every tensor. Protected tensors remain at higher precision, so the exact measured BPW is authoritative. In particular, a 6bit-named mixed plan may retain 4bit as its base precision while selecting 6-bit, 8-bit, or BF16 for other tensors to meet an approximately 6-BPW total budget. Protection floors can also raise a 4bit-named pack close to (or above) a 6bit budget on small or heavily protected models.

See the AutomatosX MLX model catalog for related MLX and OptiQ alternatives.

Download

python -m pip install -U huggingface_hub
hf download AutomatosX/AX-Qwen3.6-35B-A3B-MLX-AXQ-4bit-MTP --local-dir ./AX-Qwen3.6-35B-A3B-MLX-AXQ-4bit-MTP

Allow at least 23.12 GB of free disk space. Pin the resulting Hub commit in reproducible deployments rather than relying indefinitely on main.

Run with MLX-LM

python -m pip install -U mlx-lm
mlx_lm.generate \
  --model AutomatosX/AX-Qwen3.6-35B-A3B-MLX-AXQ-4bit-MTP \
  --prompt "Explain mixed-precision quantization in three sentences." \
  --max-tokens 128 \
  --temp 0.0

MLX-LM compatibility covers standard text/backbone inference. It may ignore AXQuant runtime metadata and optional sidecars (vision.safetensors, mtp.safetensors); this command therefore does not establish MTP acceleration or vision-language quality. The artifact records MLX 0.32.0 and MLX-LM 0.31.3 from conversion.

AX Engine status

AX Engine on df-macbookpro-m5 loads this checkpoint for Tier 1 certification. Product default remains direct fallback. Do not claim MTP acceleration for this MoE pack until Tier 2 speed gates pass. MoE MTP load requires an engine that accepts mlp.experts.gate_up_proj packing (see ax-engine PR for the load fix).

Use the packaged Qwen MTP head with oMLX or MTPLX

Download the complete repository to a writable local directory. In oMLX 0.6.3rc2 or newer, add that directory, open Model Settings, choose Import MTP side-car, and then enable Lightning MTP. The import changes only the local copy so the sidecar tensors become visible through the checkpoint index. This is a text-path compatibility result; it does not certify VLM loading or vision quality.

MTPLX can consume the packaged sidecar directly:

mtplx quickstart \
  --model ./AX-Qwen3.6-35B-A3B-MLX-AXQ-4bit-MTP \
  --profile stable \
  --depth 1 \
  --reasoning off

mtplx_runtime.json declares the canonical qwen3-next-mtp execution contract. This enables strict runtime discovery; it does not establish oMLX or MTPLX exactness or speed certification.

Quantization layout

• Quantization methods: affine, bf16.
• Group sizes used by quantized assignments: 32, 64.
• MTP sidecar: 19 tensors, 844.64M parameters, 1.69 GB, BF16.
• Vision sidecar: 333 tensors, 446.57M parameters, 0.89 GB, BF16.
• Optimization scope: text-path.
• Support tier: convertible.

BF16 sidecars, when present, are included in total download size. Their presence does not by itself establish MTP acceleration or vision-language quality.

Evidence and validation status

Modalities (capability-gated)

Text checkpoint Tier 1 does not imply vision or audio quality. Vision present=true on a pack is not a quality pass.

Intended use and limitations

• Intended for local development and evaluation on Apple Silicon with MLX-compatible runtimes.

• No minimum unified-memory figure is claimed; loadability depends on model size, context length, KV-cache policy, runtime buffers, and other processes using unified memory.

• Architecture-prior allocation is not measured sensitivity. It must not be presented as measured model quality.

• MTP requires a sidecar-aware runtime. oMLX/MTPLX discovery compatibility does not establish exactness or speed certification for those runtimes.

• Vision weights are byte-preserved at BF16, but this release does not claim validated VLM quality.

• The configured context window can require substantially more memory as the KV cache grows.

• AX Engine execution is not established because this package has no validated native manifest.

• Upstream capabilities, limitations, biases, and responsible-use guidance still apply.

Provenance and audit files

• axquant_manifest.json: package identity, byte accounting, runtime contract, software versions, and file checksums.
• axquant_plan.json: per-tensor precision decisions and planning evidence.
• axquant_quantizer_execution.json: conversion coverage and fallback records.
• axquant_runtime.json: declared AX Engine and MLX-LM compatibility metadata; runtime checks remain separate evidence.
• axquant_mtp_sidecar_manifest.json: MTP tensor provenance.
• axquant_vision_sidecar_manifest.json: protected vision tensor provenance.

All published provenance uses repository-relative paths. Local source paths are stripped before publication. The checkpoint was converted from BF16 rather than re-quantized from an OptiQ artifact. Parallel OptiQ repositories use a different quantizer and should not be assumed to have identical BPW or quality.

License

The checkpoint follows the upstream model license where applicable (often Apache License 2.0). See the Qwen/Qwen3.6-35B-A3B model card for license terms, model limitations, and responsible-use guidance.