What are the hardware requirements for Qwen3.5 397B A17B 4bit?

Qwen3.5 397B A17B 4bit requires approximately 222.2 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwen3.5 397B A17B 4bit and how large is it?

Qwen3.5 397B A17B 4bit is developed by mlx-community, a model with 397b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Qwen3.5 397B A17B 4bit perform on standard benchmarks?

Qwen3.5 397B A17B 4bit has the following published scores: LMSYS ELO 1442. Compare against reference models on this page or on the LLM Explorer leaderboards.

Qwen3.5 397B A17B 4bit by mlx-community — VRAM 222.2GB

Name: Qwen3.5 397B A17B 4bit
Author: mlx-community

Qwen3.5 397B A17B 4bit is an open-source language model by mlx-community. Features: 397b LLM, VRAM: 222.2GB, License: apache-2.0, Quantized, LLM Explorer Score: 0.42, ELO: 1442.

4-bit 4bit Conversational Endpoints compatible Image-text-to-text Mlx Quantized Qwen3 5 moe Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/mlx-community/Qwen3.5-397B-A17B-4bit

Qwen3.5 397B A17B 4bit Benchmarks

LMSys ELO 1442 vs 1272 (so35) 13.4%

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Qwen3.5 397B A17B 4bit Parameters and Internals

LLM Name	Qwen3.5 397B A17B 4bit
Repository 🤗	https://huggingface.co/mlx-community/Qwen3.5-397B-A17B-4bit
Model Size	397b
Required VRAM	222.2 GB
Updated	2026-08-08
Maintainer	mlx-community
Model Type	qwen3_5_moe
Model Files	5.2 GB: 1-of-46 4.9 GB: 2-of-46 4.9 GB: 3-of-46 5.0 GB: 4-of-46 4.9 GB: 5-of-46 4.9 GB: 6-of-46 5.0 GB: 7-of-46 4.9 GB: 8-of-46 4.9 GB: 9-of-46 5.0 GB: 10-of-46 4.9 GB: 11-of-46 4.9 GB: 12-of-46 5.0 GB: 13-of-46 4.9 GB: 14-of-46 4.9 GB: 15-of-46 5.0 GB: 16-of-46 4.9 GB: 17-of-46 4.9 GB: 18-of-46 5.0 GB: 19-of-46 4.9 GB: 20-of-46 4.9 GB: 21-of-46 5.0 GB: 22-of-46 4.9 GB: 23-of-46 4.9 GB: 24-of-46 5.0 GB: 25-of-46 4.9 GB: 26-of-46 4.9 GB: 27-of-46 5.0 GB: 28-of-46 4.9 GB: 29-of-46 4.9 GB: 30-of-46 5.0 GB: 31-of-46 4.9 GB: 32-of-46 4.9 GB: 33-of-46 5.0 GB: 34-of-46 4.9 GB: 35-of-46 4.9 GB: 36-of-46 5.0 GB: 37-of-46 4.9 GB: 38-of-46 4.9 GB: 39-of-46 5.0 GB: 40-of-46 4.9 GB: 41-of-46 4.9 GB: 42-of-46 5.0 GB: 43-of-46 4.9 GB: 44-of-46 4.9 GB: 45-of-46
Quantization Type	4bit
Model Architecture	Qwen3_5MoeForConditionalGeneration
License	apache-2.0
Transformers Version	4.57.0.dev0

Rank the Qwen3.5 397B A17B 4bit 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 Qwen3.5 397B A17B 4bit

Best Alternatives	Context / RAM	Downloads	Likes
Qwen3.5 397B A17B 8bit	0K / 209.7 GB	105802	5
Ornith 1.0 397B Mlx 8bit	0K / 208.8 GB	855	3
Qwen3.5 397B A17B 6bit	0K / 238.5 GB	2598	2
Qwen3.5 397B A17B	0K / 369.8 GB	331990	1549
Qwen3.5 397B A17B NVFP4	0K / 251.1 GB	428565	102
Qwen3.5 397B A17B GGUF	0K / 0.9 GB	63434	253
Qwen3.5 397B A17B NVFP4 V2	0K / 243.5 GB	23781	5
Ornith 1.0 397B	0K / 288.5 GB	318396	260
Ornith 1.0 397B FP8	0K / 147.7 GB	597119	183
XYZ Aquila Pro	0K / 225 GB	1805	375

Note: green Score (e.g. "73.2") means that the model is better than mlx-community/Qwen3.5-397B-A17B-4bit.