LLM EXPLORER 58,531 MODELS INDEXED

Qwen3 30B A3B MLX 8bit by Qwen

By Qwen · 80 downloads
HF Repo ↗ Copy URL

Qwen3 30B A3B MLX 8bit is an open-source language model by Qwen. Features: 30b LLM, VRAM: 31.4GB, Context: 40K, License: apache-2.0, Quantized, LLM Explorer Score: 0.19.

  Arxiv:2309.00071   Arxiv:2505.09388   8-bit   8bit   Conversational   Mlx   Quantized   Qwen3 moe   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/Qwen/Qwen3-30B-A3B-MLX-8bit 

Qwen3 30B A3B MLX 8bit Parameters and Internals

LLM NameQwen3 30B A3B MLX 8bit
Repository 🤗https://huggingface.co/Qwen/Qwen3-30B-A3B-MLX-8bit 
Model Size30b
Required VRAM31.4 GB
Updated2026-06-20
MaintainerQwen
Model Typeqwen3_moe
Model Files  5.2 GB: 1-of-6   5.4 GB: 2-of-6   5.3 GB: 3-of-6   5.3 GB: 4-of-6   5.4 GB: 5-of-6   4.8 GB: 6-of-6
Quantization Type8bit
Model ArchitectureQwen3MoeForCausalLM
Licenseapache-2.0
Context Length40960
Model Max Length40960
Transformers Version4.51.0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151936
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Qwen3 30B A3B MLX 8bit

Best Alternatives
Context / RAM
Downloads
Likes
...en3 30B A3B Instruct 2507 4bit256K / 17.2 GB8376612
...30B A3B Instruct 2507 MLX 4bit256K / 17.2 GB246528
...30B A3B Instruct 2507 MLX 8bit256K / 32.5 GB246553
...n3 Coder 30B A3B Instruct 4bit256K / 17.2 GB725233
Unslopper 30B A3B 6bit256K / 24.7 GB721
...r 30B A3B Instruct 4bit Dwq V2256K / 17.2 GB134210
...n3 Coder 30B A3B Instruct 8bit256K / 32.5 GB12777
...oder 30B A3B Instruct 4bit DWQ256K / 17.2 GB6445
...30B A3B Thinking 2507 MLX 4bit256K / 17.2 GB5761
...en3 30B A3B Thinking 2507 4bit256K / 17.2 GB2324
Note: green Score (e.g. "73.2") means that the model is better than Qwen/Qwen3-30B-A3B-MLX-8bit.