Skip to main content
Watch: I Built an AI App Without Leaving Databricks

Meta - Mixture of experts - verified

Run Llama 4 Maverick 17B-128E Instruct locally

Llama 4 Maverick 17B-128E Instruct is a 401.6B mixture-of-experts (17B active) model from Meta with a 1M-token context window, released under Llama 4 Community License Agreement (commercial use is allowed with conditions set out in the license); the smallest listed download is 242.8 GB, the Q4_K_M GGUF build from unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF.

Official weights: meta-llama/Llama-4-Maverick-17B-128E-Instruct - modalities: text, image

Total params
401.6B
Active params
17B
Context
1M
Smallest download
242.8 GB

Fits in

Smallest download at most 85% of memory

None of the tiers here. The smallest listed download is 242.8 GB, more than 85% of 256 GB.

Downloads

The official weights and every quantization listed for this model, with file sizes from Hugging Face.

Llama 4 Maverick 17B-128E Instruct downloads and file sizes
DownloadFormatSizeRepo
Official weightsBF16803.2 GBmeta-llama/Llama-4-Maverick-17B-128E-Instruct
Q4_K_MSmallest listedGGUF242.8 GBunsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF
Q8_0GGUF425.8 GBunsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF

GB is decimal: 1 GB = 10^9 bytes. Add memory for the context and the operating system on top of the file size.

Run it

Copied verbatim from the page linked under each command. Server commands from model cards often assume a multi-GPU node; check the flags before you run them.

Ollama

Terminal
ollama run llama4:maverick

Source: ollama.com/library/llama4

Transformers

Terminal
pip install -U transformers

Source: huggingface.co/meta-llama/Llama-4-Scout-17B-16E-Instruct

Not sure which runtime to use? Compare local LLM runtimes by platform, API and features.

License

Commercial use
Commercial use: ConditionsCommercial use is allowed with conditions set out in the license.
Gated download
Yes. Hugging Face asks you to accept the terms before downloading.

Worth knowing

Gated download under the Llama 4 Community License, and even the Q4_K_M build is about 243GB, so it is workstation or multi-GPU territory.

Released (per the model card).

FAQ

How big is the Llama 4 Maverick 17B-128E Instruct download?

The official BF16 weights in meta-llama/Llama-4-Maverick-17B-128E-Instruct are 803.2 GB. The smallest listed download is 242.8 GB, the Q4_K_M GGUF build from unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF. Plan for more memory than the file size: the context (KV cache) and the operating system need room too.

Can I use Llama 4 Maverick 17B-128E Instruct commercially?

Llama 4 Maverick 17B-128E Instruct is released under Llama 4 Community License Agreement, and commercial use is allowed with conditions set out in the license. The Hugging Face download is gated: you accept the license terms before it starts. The license text is at https://github.com/meta-llama/llama-models/blob/main/models/llama4/LICENSE.

How much memory do I need for Llama 4 Maverick 17B-128E Instruct?

Its smallest listed download is 242.8 GB, more than 85% of 256 GB, so it needs a multi-GPU server or a machine with more memory than the largest tier here.

How do I run Llama 4 Maverick 17B-128E Instruct with Ollama?

Run "ollama run llama4:maverick", as shown at https://ollama.com/library/llama4. The Ollama library lists the llama4:maverick tag at 245GB.

What is the context length of Llama 4 Maverick 17B-128E Instruct?

1,000,000 tokens. A longer context needs more memory for the KV cache, so you can run with a shorter one when memory is tight.

Sources

Last verified from the pages below.

Newsletter

New open weights, every week

One email per week. New model releases, open-weight drops and what they mean for what you build.

Keep exploring

Best local AI models

Our curated picks by hardware tier, with benchmarks and the reasoning.

Model API pricing and specs

First-party prices per 1M tokens when you would rather call an API.

AI benchmark leaderboards

Sourced scores, for when you need to compare quality.

Setup guides

Point your coding tools at the model you picked.