AX-DeepSeek-V4-Flash-MLX-AXQ-2bit

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 are preserved at BF16 in the checkpoint (or a bound sidecar when present).

Development evidence — not a certified AXQuant release. This package has conversion and artifact-integrity records, but it does not publish measured quality, long-context, kernel-speed, or MTP-speed evidence. Do not interpret the AXQ product label as a benchmark claim.

Model details

Property	Value
Base model	deepseek-ai/DeepSeek-V4-Flash
Source revision	60d8d70770c6776ff598c94bb586a859a38244f1
Product family	deepseek-v4
Source architecture	DeepseekV4ForCausalLM (mixture of experts (MoE)); text path optimized
Main-model parameters	284.33B logical parameters
Quantizer	AXQuant 1.5.1
Hub budget class	2bit
AXQuant base precision class	2bit-experimental
Planned storage-adjusted BPW	3.4232
Measured main-model BPW	3.1329
Measured total BPW, including MTP	3.1605
Safetensors weight size	114.94 GB
Approximate complete download	115.02 GB
Configured maximum context	1,048,576 tokens; practical limits depend on unified memory
Primary MLX runtime	MLX-LM
AX Engine native execution	Not established; no validated native manifest is included
MTP present	True
Vision present	False
Audio present	False

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.

Sibling	Intended trade-off
4bit sibling	Lower-storage AXQ budget; check its exact BPW
4bit sibling	Higher average precision near the 4-BPW budget

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

Download

python -m pip install -U huggingface_hub
hf download AutomatosX/AX-DeepSeek-V4-Flash-MLX-AXQ-2bit --local-dir ./AX-DeepSeek-V4-Flash-MLX-AXQ-2bit

Allow at least 115.02 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-DeepSeek-V4-Flash-MLX-AXQ-2bit \
  --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

This package does not include a validated native model-manifest.json, so AX Engine execution is not established by this release. The AX Engine fields in axquant_runtime.json describe the intended compatibility contract, not observed runtime evidence. Use the architecture-specific MLX runtime path above. The artifact records AX Engine version 6.11.1, but version discovery alone is not a runtime check.

Quantization layout

Main-weight precision	Parameters	Share
2bit	278.11B	95.59%
4bit	3.64B	1.25%
8bit	529.53M	0.18%
bf16	8.67B	2.98%

• Quantization methods: affine, bf16.
• Group sizes used by quantized assignments: 32.
• MTP sidecar: 1575 tensors, 6.61B parameters, 3.59 GB, BF16, F32, F8_E4M3, F8_E8M0, I8.
• Vision sidecar: not included.
• 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

Check	Status
Planning evidence	architecture_prior
Calibration	none; the allocation is based on architecture priors
Quantizer execution	33492/33492 recorded module conversions succeeded; 0 fallbacks
AX Engine native manifest	not included
Quality versus BF16 or uniform baselines	Not published; no quality-retention claim
MTP acceptance and speed	not measured; no MTP speedup claim
AX Engine kernel evidence	unmeasured
Vision-language quality	Not applicable (no vision tower in this package)
Speech-recognition quality	Not applicable
Long-context quality	1,048,576-token capacity is config metadata, not a validated claim
Release certification	Not certified; formal AXQuant M0-M8 gates are not closed

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 may be ignored outside AX Engine and its speedup is unmeasured for this exact checkpoint.

• 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 compatibility metadata; runtime checks remain separate evidence.
• axquant_mtp_sidecar_manifest.json: MTP 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. 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 deepseek-ai/DeepSeek-V4-Flash model card for license terms, model limitations, and responsible-use guidance.