mlx-optiq
← Models

Mage-VL-OptiQ-4bit

microsoft/Mage-VL, quantized by OptiQ for Apple Silicon; takes images.

vision
On disk3.9 GB
Memoryabout 6 GB
Capability Score70.27
Imagesyes

Capability Score breakdown

The six metrics behind this model's Capability Score of 70.27, as reported in the model card.

Capability Score breakdown: Mage-VL-OptiQ-4bitMMLU: This model 74.6; GSM8K: This model 88.7; IFEval: This model 68.6; BFCL-V3: This model 88.5; HumanEval: This model 76.2; HashHop: This model 25.0MMLU74.6GSM8K88.7IFEval68.6BFCL-V388.5HumanEval76.2HashHop25.0

Use it

Install

pip install mlx-optiq

Serve it

An OpenAI- and Anthropic-compatible server on this Mac. Every app below can use it.

optiq serve --model mlx-community/Mage-VL-OptiQ-4bit

OptiQ Code

The terminal coding agent uses whatever model optiq serve is running.

optiq serve --model mlx-community/Mage-VL-OptiQ-4bit
optiq code

OptiQ Lab

optiq lab --model mlx-community/Mage-VL-OptiQ-4bit

Claude Code

optiq run claude -m mlx-community/Mage-VL-OptiQ-4bit

Codex

optiq run codex -m mlx-community/Mage-VL-OptiQ-4bit

OpenCode

optiq run opencode -m mlx-community/Mage-VL-OptiQ-4bit

Mistral Vibe

optiq run vibe -m mlx-community/Mage-VL-OptiQ-4bit

Hermes Agent

optiq run hermes -m mlx-community/Mage-VL-OptiQ-4bit

OpenClaw

optiq run openclaw -m mlx-community/Mage-VL-OptiQ-4bit

optiq run starts optiq serve if it is not running, sets the agent up for this session only, and stops the server when the agent exits.

mlx-lm, text only

from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Mage-VL-OptiQ-4bit")
print(generate(model, tokenizer, prompt="Hello", max_tokens=256))

Details

FamilyMage-VL · Mage-VL docs
ArchitectureMage-VL
Base modelmicrosoft/Mage-VL
Size on disk3.9 GB
Memory neededabout 6 GB
Imagesyes, through optiq serve and the Lab
Capability Score70.27 (from the model card)
Downloads208 in the last month
Updated2026-09-15
Hugging Facemlx-community/Mage-VL-OptiQ-4bit