Llama 4 Maverick 17B 128E Instruct (Q5_K_S) — 49.3 GBon Apple M2 Pro 16GB
Overview
Llama 4 Maverick 17B 128E Instruct is a 396.58B parameter moe language model by Meta, with code, multilingual, tool-calls, vision capabilities. It supports a context window of up to 1,048,576 tokens.
Llama 4 Maverick 17B 128E Instruct is a large-scale Mixture-of-Experts model from Meta with 17 billion parameters per expert and 128 experts, activating one expert per token for a total of approximately 400 billion parameters. It delivers frontier-class performance on vision, code generation, and multilingual tasks across 12 languages. Maverick represents the high-capacity tier of the Llama 4 family, trading higher memory requirements for stronger benchmark results. With a 1M token context window, it requires multi-GPU setups but quantizes down to Q2 levels.
At Q5_K_S quantization (medium quality tier), the model weighs 256.77 GB. This exceeds the 16 GB of VRAM on Apple M2 Pro 16GB. Inference is still possible via CPU offload or memory-mapped loading from disk, but expect significantly reduced performance.
Hardware Requirements
| Model size | 256.77 GB |
| VRAM available | 16 GB |
| VRAM used | 49.3 GB |
| System RAM | |
| Min RAM required | 256.8 GB |
| GPU layers | 0 / 48 |
| Context size | 262,144 |
| Backend | metal |
| Flash attention | No |
| Reading from disk | Yes |
Performance Notes
Deploy
Install llama.cpp
brew install llama.cppDownload Model
curl -L -o llama-4-maverick-17b-128e-instruct.gguf "https://huggingface.co/unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF/resolve/main/Q5_K_S/Llama-4-Maverick-17B-128E-Instruct-Q5_K_S-00001-of-00006.gguf"Start Server
llama-server \
-m llama-4-maverick-17b-128e-instruct.gguf \
--n-gpu-layers 0 \
--ctx-size 262144Verify
curl http://localhost:8080/healthFrequently Asked Questions
How much VRAM does Llama 4 Maverick 17B 128E Instruct (Q5_K_S) need?
The Q5_K_S quantization of Llama 4 Maverick 17B 128E Instruct requires 256.77 GB. The 16 GB of VRAM on Apple M2 Pro 16GB is insufficient for GPU layers, so inference runs on CPU.
Can I run Llama 4 Maverick 17B 128E Instruct on Apple M2 Pro 16GB?
It is possible but not recommended. Apple M2 Pro 16GB does not have enough VRAM to accelerate Llama 4 Maverick 17B 128E Instruct (Q5_K_S), so inference will rely on CPU and system RAM.
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. Q5_K_S compresses Llama 4 Maverick 17B 128E Instruct from its original size down to 256.77 GB.
What quantization should I choose for Llama 4 Maverick 17B 128E Instruct?
Q5_K_S 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.
Why are some layers offloaded to CPU?
Apple M2 Pro 16GB has 16 GB of VRAM, but Llama 4 Maverick 17B 128E Instruct (Q5_K_S) requires approximately 256.77 GB. Only 0 of 48 layers fit in VRAM; the remaining layers run on CPU, which is slower but still functional.
What is MoE and how does it affect deployment?
Llama 4 Maverick 17B 128E Instruct uses a Mixture-of-Experts (MoE) architecture with 128 experts, of which 1 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 Llama 4 Maverick 17B 128E Instruct (Q5_K_S) with Ollama?
Run ollama run llama4:17b-maverick-128e-instruct-q5_k_s to start Llama 4 Maverick 17B 128E Instruct (Q5_K_S). Ollama handles downloading the model weights automatically on first run.