Instructions to use AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP 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("AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP")

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 AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP 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 "AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP"

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 AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP"

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 "AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP" \
  --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"

AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP

Checkpoint Tier 1 certified on df-macbookpro-m5 (2026-08-09) for Hub commit 85b0a78a1484. Measured size vs matched uniform baseline, quality retention ≥0.98, and conversion integrity. Tier 1 is not a speed claim: MTP acceleration is not certified. See the certificate.

AXQuant mixed-precision (AXQ) Gemma 4 target with multi-token prediction (MTP) — named like the Qwen AXQ-MTP line (…-MLX-AXQ-*-MTP).

Property	Value
Product class	AXQ 4bit
Target size	26B-A4B
Measured BPW	~4.90
Hub id	`AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP`
Hub commit	`85b0a78a14843a818d403f9a2525efa2f081c2a4`
Engine pair	`gemma-4-26b-a4b-it` ↔ `gemma-4-26b-a4b-it-assistant`
MTP layout	`assistant/` + `ax_gemma4_assistant_mtp.json` (Gemma assistant-MTP)
Certification host	`df-macbookpro-m5`

Layout

./                          # AXQ target weights (+ vision sidecar when present)
assistant/                  # gemma4_assistant drafter
ax_gemma4_assistant_mtp.json
ax_composite_pack_manifest.json
axquant_*.json

Naming follows Qwen AXQ MTP packs: the -MTP suffix means MTP assets are in-repo.

Certification

Claim	Status
Checkpoint Tier 1 (size + quality + load)	Certified on `df-macbookpro-m5`
MTP acceleration Tier 2	Not certified (assistant-MTP is present for product completeness only)
Vision / multimodal quality	Not claimed

Certificate: gemma4-26b-a4b-axq4-tier1.md

Runtime (AX Engine 7.1.5)

Download the complete snapshot and serve its local root:

hf download AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP --local-dir ./AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP
ax-engine serve ./AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP --port 31418

AX Engine validates ax_gemma4_assistant_mtp.json and the exact-paired assistant/ bundle before attaching the drafter. Assistant-MTP is enabled by default in AX Engine 7.1.5 with a maximum draft depth of two. Set AX_MLX_GEMMA4_ASSISTANT_MTP=0 to force direct decode, or set AX_MLX_GEMMA4_ASSISTANT_MTP_MAX_DEPTH=1 to cap drafting at one token.

Stock MLX-LM does not consume the assistant/ bundle. This is not a Qwen sidecar, so the oMLX/MTPLX Qwen import workflow does not apply. The checkpoint card still makes no Tier 2 MTP acceleration claim.

Published / card updated: 2026-08-21.

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.

Modality	Claim	Supported	Reason
Vision	`present-not-certified`	`true`	vision present sidecar=['vision.safetensors']; mlx-vlm smoke failed on df-macstudio-m2 (mlx-vlm expects vision_tower.*; sidecar/layout mismatch). Text Tier 1 unchanged. Evidence: docs/certifications/evidence/modality-recert-capability-gated/results/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP.json
Audio	`not-applicable`	`false`	audio not supported (no tower config and no sidecar weights)

Downloads last month: 73

Safetensors

Model size

4B params

Tensor type

BF16

U32

MLX

Hardware compatibility

4-bit

Collections including AutomatosX/AX-gemma-4-26b-a4b-MLX-AXQ-4bit-MTP