qwen36-35b-a3b-uncensored-optiq-4b-mtp

Qwen/Qwen3.6-35B-A3B, quantized with OptiQ by Ted17846578368756 for Apple Silicon; mixture-of-experts, takes images.

communityvisionMoEMTP head

On disk24.7 GB

Memoryabout 28 GB

Imagesyes

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 Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp

OptiQ Code

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

optiq serve --model Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp
optiq code

OptiQ Lab

optiq lab --model Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp

Claude Code

optiq run claude -m Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp

Codex

optiq run codex -m Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp

OpenCode

optiq run opencode -m Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp

Mistral Vibe

optiq run vibe -m Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp

Hermes Agent

optiq run hermes -m Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp

OpenClaw

optiq run openclaw -m Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp

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("Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp")
print(generate(model, tokenizer, prompt="Hello", max_tokens=256))

Details

Family	Qwen · Qwen docs
Architecture	Qwen3.5 MoE
Base model	Qwen/Qwen3.6-35B-A3B
Size on disk	24.7 GB
Memory needed	about 28 GB
Images	yes, through optiq serve and the Lab
MTP head	included
Downloads	119 in the last month
Updated	2026-09-20
Hugging Face	Ted17846578368756/qwen36-35b-a3b-uncensored-optiq-4b-mtp