LLM EXPLORER 57,918 MODELS INDEXED

Qwen3.6 35B A3B UD MLX 3bit by unsloth

By unsloth · 2292 downloads
HF Repo ↗ Copy URL

Qwen3.6 35B A3B UD MLX 3bit is an open-source language model by unsloth. Features: 35b LLM, VRAM: 17.3GB, License: apache-2.0, Quantized, LLM Explorer Score: 0.42, ELO: 1395.

  3-bit   3bit Base model:quantized:qwen/qwen... Base model:qwen/qwen3.6-35b-a3...   Conversational   Image-text-to-text   Mlx   Quantized   Qwen   Qwen3 5 moe   Region:us   Safetensors   Sharded   Tensorflow   Unsloth
Model Card on HF 🤗: https://huggingface.co/unsloth/Qwen3.6-35B-A3B-UD-MLX-3bit 

Qwen3.6 35B A3B UD MLX 3bit Benchmarks

LMSys ELO 1395 vs 1272 (so35) 9.7%
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

Qwen3.6 35B A3B UD MLX 3bit Parameters and Internals

LLM NameQwen3.6 35B A3B UD MLX 3bit
Repository 🤗https://huggingface.co/unsloth/Qwen3.6-35B-A3B-UD-MLX-3bit 
Base Model(s)  Qwen/Qwen3.6-35B-A3B   Qwen/Qwen3.6-35B-A3B
Model Size35b
Required VRAM17.3 GB
Updated2026-08-10
Maintainerunsloth
Model Typeqwen3_5_moe
Model Files  5.3 GB: 1-of-4   5.3 GB: 2-of-4   5.3 GB: 3-of-4   1.4 GB: 4-of-4
Quantization Type3bit
Model ArchitectureQwen3_5MoeForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version4.57.1
Tokenizer ClassTokenizersBackend
Padding Token<|vision_pad|>
Errorsreplace

Best Alternatives to Qwen3.6 35B A3B UD MLX 3bit

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3.6 35B A3B MLX 4bit0K / 20.5 GB5131936
Qwen3.6 35B A3B MLX 6bit0K / 28.9 GB4785922
Qwen3.6 35B A3B UD MLX 4bit0K / 21.6 GB15720112
Qwen3.6 35B A3B MLX 8bit0K / 37.8 GB386241
Qwen3.6 35B A3B 6bit0K / 28.9 GB27428
Qwen3.6 35B A3B 5bit0K / 24.7 GB16285
Qwen3.6 35B A3B 4bit0K / 20.5 GB6203692
Qwen3.6 35B A3B 8bit0K / 37.8 GB12017527
Ornith 1.0 35B 4bit0K / 20.5 GB683513
Qwen3.6 35B A3B OptiQ 4bit0K / 22.1 GB1473044
Note: green Score (e.g. "73.2") means that the model is better than unsloth/Qwen3.6-35B-A3B-UD-MLX-3bit.