mlx-optiq
← Models

Laguna-XS-2.1-OptiQ-4bit

poolside/Laguna-XS-2.1, quantized by OptiQ for Apple Silicon.

On disk21.0 GB
Memoryabout 26 GB
Capability Score85.70

Capability Score breakdown

The six metrics behind this model's Capability Score of 85.70, as reported in the model card.

Capability Score breakdown: Laguna-XS-2.1-OptiQ-4bitMMLU: This model 86.2; GSM8K: This model 95.6; IFEval: This model 76.5; BFCL-V3: This model 89.0; HumanEval: This model 82.9; HashHop: This model 84.0MMLU86.2GSM8K95.6IFEval76.5BFCL-V389.0HumanEval82.9HashHop84.0

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/Laguna-XS-2.1-OptiQ-4bit

OptiQ Code

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

optiq serve --model mlx-community/Laguna-XS-2.1-OptiQ-4bit
optiq code

OptiQ Lab

optiq lab --model mlx-community/Laguna-XS-2.1-OptiQ-4bit

Claude Code

optiq run claude -m mlx-community/Laguna-XS-2.1-OptiQ-4bit

Codex

optiq run codex -m mlx-community/Laguna-XS-2.1-OptiQ-4bit

OpenCode

optiq run opencode -m mlx-community/Laguna-XS-2.1-OptiQ-4bit

Mistral Vibe

optiq run vibe -m mlx-community/Laguna-XS-2.1-OptiQ-4bit

Hermes Agent

optiq run hermes -m mlx-community/Laguna-XS-2.1-OptiQ-4bit

OpenClaw

optiq run openclaw -m mlx-community/Laguna-XS-2.1-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/Laguna-XS-2.1-OptiQ-4bit")
print(generate(model, tokenizer, prompt="Hello", max_tokens=256))

Details

FamilyLaguna · Laguna docs
ArchitectureLaguna
Base modelpoolside/Laguna-XS-2.1
Size on disk21.0 GB
Memory neededabout 26 GB
Imagesno, text only
Capability Score85.70 (from the model card)
Downloads373 in the last month
Updated2026-10-16
Hugging Facemlx-community/Laguna-XS-2.1-OptiQ-4bit