← Models

Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

mlx-community/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B, quantized with OptiQ by vadimayegorov for Apple Silicon; mixture-of-experts.

communityMoEexpert-pruned

On disk12.3 GB

Memoryabout 16 GB

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 vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

OptiQ Code

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

optiq serve --model vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16
optiq code

OptiQ Lab

optiq lab --model vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

Claude Code

optiq run claude -m vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

Codex

optiq run codex -m vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

OpenCode

optiq run opencode -m vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

Mistral Vibe

optiq run vibe -m vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

Hermes Agent

optiq run hermes -m vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

OpenClaw

optiq run openclaw -m vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16

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("vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16")
print(generate(model, tokenizer, prompt="Hello", max_tokens=256))

Details

Family	Qwen · Qwen docs
Architecture	Qwen3.5 MoE
Base model	mlx-community/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B
Size on disk	12.3 GB
Memory needed	about 16 GB
Images	no, text only
Downloads	178 in the last month
Updated	2026-09-03
Hugging Face	vadimayegorov/Qwen3.6-35B-A3B-OptiQ-4bit-REAP-19B-mlx-fp16