Qwen3.5 35B A3B (Q4_K_XL) — 30.5 GBon Apple M3 Max 64GB
Overview
Qwen3.5 35B A3B is a 35.95B parameter moe language model by Qwen, with code, multilingual, thinking, tool-calls, vision capabilities. It supports a context window of up to 262,144 tokens.
Qwen3.5 35B A3B is a Mixture-of-Experts model from Alibaba's Qwen team with 35 billion total parameters but only 3 billion active per token, routed across 256 experts for extreme efficiency. It is natively multimodal, processing text, images, and video, with built-in thinking capabilities for chain-of-thought reasoning. The model supports a 262K context window and covers over 200 languages. Released under the Apache 2.0 license, it delivers flagship-level performance at a fraction of the compute cost, quantizing efficiently for self-hosted deployment on consumer hardware.
At Q4_K_XL quantization (medium quality tier), the model weighs 19.17 GB. This fits within the 64 GB of VRAM on Apple M3 Max 64GB, enabling full GPU inference.
Hardware Requirements
| Model size | 19.17 GB |
| VRAM available | 64 GB |
| VRAM used | 30.5 GB |
| System RAM | |
| Min RAM required | 0 GB |
| GPU layers | 40 / 40 |
| Context size | 262,144 |
| Backend | metal |
| Flash attention | Yes |
| Reading from disk | Yes |
Performance Notes
Deploy
Install llama.cpp
brew install llama.cppDownload Model
curl -L -o qwen3-5-35b-a3b.gguf "https://huggingface.co/unsloth/Qwen3.5-35B-A3B-GGUF/resolve/main/Qwen3.5-35B-A3B-UD-Q4_K_XL.gguf"Start Server
llama-server \
-m qwen3-5-35b-a3b.gguf \
--n-gpu-layers 40 \
--ctx-size 262144 \
--flash-attnVerify
curl http://localhost:8080/healthFrequently Asked Questions
How much VRAM does Qwen3.5 35B A3B (Q4_K_XL) need?
The Q4_K_XL quantization of Qwen3.5 35B A3B requires 19.17 GB. All 40 layers fit in the 64 GB of VRAM available on Apple M3 Max 64GB, enabling full GPU acceleration.
Can I run Qwen3.5 35B A3B on Apple M3 Max 64GB?
Yes. Apple M3 Max 64GB provides 64 GB of VRAM, which is enough to run Qwen3.5 35B A3B (Q4_K_XL) with all layers on the GPU for optimal performance.
What is quantization?
Quantization reduces a model's numerical precision from its original floating-point format to a more compact representation. This shrinks the file size and VRAM footprint, making it possible to run large models on consumer hardware. The trade-off is a small reduction in output quality. Q4_K_XL compresses Qwen3.5 35B A3B from its original size down to 19.17 GB.
What quantization should I choose for Qwen3.5 35B A3B?
Q4_K_XL is a medium-quality quantization. Higher-quality quants (Q8, Q6) preserve more model accuracy but need more VRAM. Lower quants (Q4, Q3, Q2) reduce VRAM usage at the cost of some quality. Choose based on your available hardware and quality requirements.
What is flash attention and why is it enabled?
Flash attention is a memory-efficient algorithm that speeds up the attention mechanism in transformer models. It reduces VRAM usage during inference by avoiding the materialisation of the full attention matrix. For Qwen3.5 35B A3B on Apple M3 Max 64GB, flash attention is enabled to maximise context length and throughput within the available 64 GB of VRAM.
What is MoE and how does it affect deployment?
Qwen3.5 35B A3B uses a Mixture-of-Experts (MoE) architecture with 256 experts, of which 8 are active per token. This means only a fraction of the model weights are used for each inference step, allowing MoE models to be larger in total parameter count while remaining efficient at inference time.
How do I run Qwen3.5 35B A3B (Q4_K_XL) with Ollama?
Run ollama run qwen3.5:35b-a3b-q4_k_xl to start Qwen3.5 35B A3B (Q4_K_XL). Ollama handles downloading the model weights automatically on first run.