mlx-optiq
← Models

humanizer-1B-OptiQ-4bit

mlx-community/MiniCPM5-1B-OptiQ-4bit, quantized by OptiQ for Apple Silicon.

On disk1.1 GB
Memoryabout 3 GB
Parent score30.28 MiniCPM5-1B-OptiQ-4bit

This model reuses the per-layer bit allocation of MiniCPM5-1B-OptiQ-4bit, whose measured Capability Score is 30.28.

The parent's Capability Score breakdown

This model has no Capability Score of its own. The chart shows the six metrics behind its parent MiniCPM5-1B-OptiQ-4bit's score of 30.28, taken from the parent's model card and compared with uniform 4-bit.

  • MiniCPM5-1B-OptiQ-4bit
  • Uniform 4-bit
The parent's Capability Score breakdown: MiniCPM5-1B-OptiQ-4bitMMLU: MiniCPM5-1B-OptiQ-4bit 52.4, Uniform 4-bit 49.0; GSM8K: MiniCPM5-1B-OptiQ-4bit 2.7, Uniform 4-bit 1.7; IFEval: MiniCPM5-1B-OptiQ-4bit 64.7, Uniform 4-bit 58.6; BFCL-V3: MiniCPM5-1B-OptiQ-4bit 0.0, Uniform 4-bit 0.0; HumanEval: MiniCPM5-1B-OptiQ-4bit 57.9, Uniform 4-bit 45.7; HashHop: MiniCPM5-1B-OptiQ-4bit 4.0, Uniform 4-bit 0.0MMLU52.449.0GSM8K2.71.7IFEval64.758.6BFCL-V30.00.0HumanEval57.945.7HashHop4.00.0

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/humanizer-1B-OptiQ-4bit

OptiQ Code

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

optiq serve --model mlx-community/humanizer-1B-OptiQ-4bit
optiq code

OptiQ Lab

optiq lab --model mlx-community/humanizer-1B-OptiQ-4bit

Claude Code

optiq run claude -m mlx-community/humanizer-1B-OptiQ-4bit

Codex

optiq run codex -m mlx-community/humanizer-1B-OptiQ-4bit

OpenCode

optiq run opencode -m mlx-community/humanizer-1B-OptiQ-4bit

Mistral Vibe

optiq run vibe -m mlx-community/humanizer-1B-OptiQ-4bit

Hermes Agent

optiq run hermes -m mlx-community/humanizer-1B-OptiQ-4bit

OpenClaw

optiq run openclaw -m mlx-community/humanizer-1B-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/humanizer-1B-OptiQ-4bit")
print(generate(model, tokenizer, prompt="Hello", max_tokens=256))

Details

FamilyMiniCPM · MiniCPM docs
ArchitectureLlama
Base modelmlx-community/MiniCPM5-1B-OptiQ-4bit
Size on disk1.1 GB
Memory neededabout 3 GB
Imagesno, text only
Family referenceThis model reuses the per-layer bit allocation of MiniCPM5-1B-OptiQ-4bit, whose measured Capability Score is 30.28.
Downloads609 in the last month
Updated2026-09-15
Hugging Facemlx-community/humanizer-1B-OptiQ-4bit