What are the hardware requirements for Qwen3 235B A22B MLX 6bit?

Qwen3 235B A22B MLX 6bit requires approximately 184.3 GB of VRAM and supports a context window of 40K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwen3 235B A22B MLX 6bit and how large is it?

Qwen3 235B A22B MLX 6bit is developed by Qwen, a model with 235b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Qwen3 235B A22B MLX 6bit?

Qwen3 235B A22B MLX 6bit is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Qwen3 235B A22B MLX 6bit by Qwen — VRAM 184.3GB, 40K context

Name: Qwen3 235B A22B MLX 6bit
Author: Qwen

Qwen3 235B A22B MLX 6bit is an open-source language model by Qwen. Features: 235b LLM, VRAM: 184.3GB, Context: 40K, License: apache-2.0, Quantized, LLM Explorer Score: 0.19.

Arxiv:2309.00071 Arxiv:2505.09388 6-bit 6bit Conversational Mlx Quantized Qwen3 moe Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Qwen/Qwen3-235B-A22B-MLX-6bit

Qwen3 235B A22B MLX 6bit Parameters and Internals

LLM Name	Qwen3 235B A22B MLX 6bit
Repository 🤗	https://huggingface.co/Qwen/Qwen3-235B-A22B-MLX-6bit
Model Size	235b
Required VRAM	184.3 GB
Updated	2026-05-06
Maintainer	Qwen
Model Type	qwen3_moe
Model Files	5.1 GB: 1-of-36 5.2 GB: 2-of-36 5.2 GB: 3-of-36 5.2 GB: 4-of-36 5.2 GB: 5-of-36 5.2 GB: 6-of-36 5.2 GB: 7-of-36 5.2 GB: 8-of-36 5.2 GB: 9-of-36 5.2 GB: 10-of-36 5.2 GB: 11-of-36 5.2 GB: 12-of-36 5.2 GB: 13-of-36 5.2 GB: 14-of-36 5.2 GB: 15-of-36 5.2 GB: 16-of-36 5.2 GB: 17-of-36 5.2 GB: 18-of-36 5.2 GB: 19-of-36 5.2 GB: 20-of-36 5.2 GB: 21-of-36 5.2 GB: 22-of-36 5.2 GB: 23-of-36 5.2 GB: 24-of-36 5.2 GB: 25-of-36 5.2 GB: 26-of-36 5.2 GB: 27-of-36 5.2 GB: 28-of-36 5.2 GB: 29-of-36 5.2 GB: 30-of-36 5.2 GB: 31-of-36 5.2 GB: 32-of-36 5.2 GB: 33-of-36 5.2 GB: 34-of-36 5.2 GB: 35-of-36 2.4 GB: 36-of-36
Quantization Type	6bit
Model Architecture	Qwen3MoeForCausalLM
License	apache-2.0
Context Length	40960
Model Max Length	40960
Transformers Version	4.51.0
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	151936
Torch Data Type	float16
Errors	replace

Rank the Qwen3 235B A22B MLX 6bit capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Qwen3 235B A22B MLX 6bit

Best Alternatives	Context / RAM	Downloads	Likes
...3 235B A22B Instruct 2507 4bit	256K / 131.4 GB	2658	3
...3 235B A22B Instruct 2507 8bit	256K / 227.7 GB	2591	2
...3 235B A22B Thinking 2507 8bit	256K / 227.7 GB	515	4
...5B A22B Thinking 2507 3bit DWQ	256K / 103 GB	204	6
...5B A22B Instruct 2507 3bit DWQ	256K / 103 GB	174	5
...5B A22B Thinking 2507 4bit DWQ	256K / 131.4 GB	65	4
...5B A22B Instruct 2507 4bit DWQ	256K / 131.4 GB	66	2
Qwen3 235B A22B 4bit	40K / 131.4 GB	2526	9
Qwen3 235B A22B 8bit	40K / 238.3 GB	2042	4
Qwen3 235B A22B MLX 4bit	40K / 125.2 GB	399	16

Note: green Score (e.g. "73.2") means that the model is better than Qwen/Qwen3-235B-A22B-MLX-6bit.