AX-Qwen3.8-27B-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 at BF16 in the checkpoint (or a bound sidecar when present).

Checkpoint Tier 1 certified on df-macbookpro-m3 (2026-08-14) at Hub commit 32f448461caf — measured size against a matched uniform baseline, quality retention, and conversion integrity. Current main preserves that revision's exact Safetensors payloads while allowing metadata-only compatibility fixes. Tier 1 is a checkpoint claim, not a general speed claim: MTP acceleration is certified for the certificate's authorizing profiles only; outside that scope there is no speedup claim. See the checkpoint Tier 1 certificate and Tier 2 MTP acceleration certificate for the bound evidence and thresholds.

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. When that collapse happens, AutomatosX does not publish a separate misleading 4bit sibling for that base.

See the AutomatosX collections for the family catalog, or the complete index.

Download

python -m pip install -U huggingface_hub
hf download AutomatosX/AX-Qwen3.8-27B-MLX-AXQ-4bit-MTP --local-dir ./AX-Qwen3.8-27B-MLX-AXQ-4bit-MTP

Allow at least 18.20 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-Qwen3.8-27B-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.

Serve with AX Engine and MTP

After installing AX Engine, download the complete repository (see AXQuant for conversion, certificates, and model-card tooling) and serve the local directory:

ax-engine serve ./AX-Qwen3.8-27B-MLX-AXQ-4bit-MTP --port 31418

AX Engine is the authority for the AXQ runtime contract and native MTP sidecar. Runtime speedup claims are limited to the authorizing profiles in the linked Tier 2 certificate. The artifact records AX Engine version 6.16.1. Native model-manifest.json status: included as model-manifest.json.

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.

MTPLX can consume the packaged sidecar directly:

mtplx quickstart \
  --model ./AX-Qwen3.8-27B-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 extend the linked AX Engine certificate to oMLX or MTPLX.

Quantization layout

• Quantization methods: affine, bf16.
• Group sizes used by quantized assignments: 64.
• MTP sidecar: 15 tensors, 424.70M parameters, 0.85 GB, BF16.
• Vision sidecar: 333 tensors, 460.73M parameters, 0.92 GB, BF16.
• Vision weights: protected BF16 sidecar.
• 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

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 extend the linked AX Engine certificate to those runtimes.

• Vision weights are 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.

• 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 compatibility metadata; runtime checks remain separate evidence.
• axquant_mtp_sidecar_manifest.json: MTP tensor provenance.
• axquant_vision_sidecar_manifest.json: protected vision tensor provenance.
• model-manifest.json: AX Engine native tensor manifest.

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. If an OptiQ repository is published separately, it uses 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/Qwen3.8-27B model card for license terms, model limitations, and responsible-use guidance.