Llama 4 Scout 17B 16E Instruct (Q8_K_XL) — 385.3 GBon Apple M3 Max 64GB

Q8_K_XL Apple M3 Max 64GB

Overview

Llama 4 Scout 17B 16E Instruct is a 17B parameter moe language model by Meta, with code, multilingual, tool-calls, vision capabilities. It supports a context window of up to 10,485,760 tokens.

Llama 4 Scout 17B 16E Instruct is a Mixture-of-Experts model from Meta with 17 billion parameters per expert and 16 experts, activating one expert per token. It supports vision, code generation, tool calling, and 12 languages, making it one of the most versatile models in the Llama 4 family. Scout targets the efficiency-focused segment, offering multimodal capabilities at lower compute cost than dense models of similar quality. Its 10M token context window is among the largest available, and it quantizes well for self-hosted multi-GPU deployments.

At Q8_K_XL quantization (high quality tier), the model weighs 119.38 GB. This exceeds the 64 GB of VRAM on Apple M3 Max 64GB. Inference is still possible via CPU offload or memory-mapped loading from disk, but expect significantly reduced performance.

Hardware Requirements

Model size	119.38 GB
VRAM available	64 GB
VRAM used	385.3 GB
Min RAM required	119.4 GB
GPU layers	0 / 48
Context size	2,097,152
Backend	metal
Flash attention	No

Performance Notes

Mixture-of-Experts

This is a Mixture-of-Experts model with 16 total experts, of which 1 are active per token. Only the active experts need fast GPU access — inactive expert weights can reside in system RAM with minimal impact on generation speed.

Read the MoE deployment guide →

Reduced performance

48 of 48 layers run on CPU. Your system needs at least 119.4 GB RAM. If system RAM is insufficient, data will be read from disk on demand, with significant performance impact.

Shorter conversations

Context size has been reduced due to memory constraints. This may affect the model's ability to handle long conversations.

Deploy

Install llama.cpp

brew install llama.cpp

Download Model

curl -L -o llama-4-scout-17b-16e-instruct.gguf "https://huggingface.co/unsloth/Llama-4-Scout-17B-16E-Instruct-GGUF/resolve/main/UD-Q8_K_XL/Llama-4-Scout-17B-16E-Instruct-UD-Q8_K_XL-00001-of-00003.gguf"

Start Server

llama-server \
  -m llama-4-scout-17b-16e-instruct.gguf \
  --n-gpu-layers 0 \
  --ctx-size 2097152

Verify

curl http://localhost:8080/health

Explore

• Configure a deployment
• Browse models
• Browse hardware

Other quantizations

• Llama 4 Scout 17B 16E Instruct (Q8_0)
• Llama 4 Scout 17B 16E Instruct (Q5_K_M)
• Llama 4 Scout 17B 16E Instruct (Q4_K_M)

Other hardware

• Llama 4 Scout 17B 16E Instruct on CPU Only
• Llama 4 Scout 17B 16E Instruct on NVIDIA A100 80GB
• Llama 4 Scout 17B 16E Instruct on NVIDIA H100

Frequently Asked Questions

How much VRAM does Llama 4 Scout 17B 16E Instruct (Q8_K_XL) need?

The Q8_K_XL quantization of Llama 4 Scout 17B 16E Instruct requires 119.38 GB. The 64 GB of VRAM on Apple M3 Max 64GB is insufficient for GPU layers, so inference runs on CPU.

Can I run Llama 4 Scout 17B 16E Instruct on Apple M3 Max 64GB?

It is possible but not recommended. Apple M3 Max 64GB does not have enough VRAM to accelerate Llama 4 Scout 17B 16E Instruct (Q8_K_XL), 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. Q8_K_XL compresses Llama 4 Scout 17B 16E Instruct from its original size down to 119.38 GB.

What quantization should I choose for Llama 4 Scout 17B 16E Instruct?

Q8_K_XL is a high-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 M3 Max 64GB has 64 GB of VRAM, but Llama 4 Scout 17B 16E Instruct (Q8_K_XL) requires approximately 119.38 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 Scout 17B 16E Instruct uses a Mixture-of-Experts (MoE) architecture with 16 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 Scout 17B 16E Instruct (Q8_K_XL) with Ollama?

Run ollama run llama4:17b-scout-16e-instruct-q8_k_xl to start Llama 4 Scout 17B 16E Instruct (Q8_K_XL). Ollama handles downloading the model weights automatically on first run.

Last updated: March 5, 2026