Skip to content
localmodel.run

Text model · Llama 4

Llama 4 Maverick: RAM and VRAM requirements

Llama 4 Maverick needs about 231.7 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~226.09 GB to download; KV cache and overhead add the rest), or about 402.2 GB at Q8_0. The lightest hardware that runs it is a high-memory machine.

Llama 4 family · 400B params (Mixture-of-Experts: activates only 17B of 400B params per token, so generation is faster than the total size suggests) · released Apr 2025.

Shopping for hardware? See what runs Llama 4 Maverick →

License Llama 4 Community License · Conditional ↓ 43.5K/mo ♥ 505 on HuggingFace
Q4_K_M GGUF
226.09 GB
Q8_0 GGUF
396.57 GB
Memory @ Q4 (4k)
~231.7 GB
Context
1000 k

Q4_K_M and Q8_0 are GGUF download (file) sizes. Memory @ Q4 includes the KV cache and overhead, so it runs larger than the file on disk.

Will it run on your device?

Llama 4 Maverick runs on 0 of 40 tracked devices at Q4_K_M.

0 run well 0 tight fit 40 too small
Fit check Q4_K_M
No, not enough memory
needs 231.7 GB usable 10.5 GB
See the full breakdown

Memory breakdown

Weights (Q4_K_M)226.09 GB
+
KV cache (4k)4.8 GB
+
Overhead0.8 GB
=
Total231.7 GB

How context length changes it

4k context ~231.7 GB 32k context ~265.3 GB 128k context ~380.5 GB

Longer context grows the KV cache, which for Llama 4 Maverick is sized by its 17B active params, not the full 400B. It needs ~231.7 GB at 4k and ~380.5 GB at 128k.

Benchmark scores

Aider polyglot · coding

#8 of 10 ranked · Percent of 225 Exercism exercises solved across six languages.

15.6%

BFCL · tool use

#6 of 19 ranked · Overall Acc on the Berkeley Function-Calling Leaderboard.

37.3%

Sourced third-party benchmarks for the full-precision Llama 4 Maverick, not a specific quant. See the full leaderboard.

Quantization sizes

GGUF quantson disk
Quantization Size on disk
Q2_K 167.5 GB est
Q3_K_M 195.5 GB est
Q4_K_M (default) 226.09 GB
Q5_K_M 285 GB est
Q6_K 328 GB est
Q8_0 396.57 GB
FP16 801 GB

Lower quant = smaller and faster, slightly lower quality. Q4_K_M is the common default.

Run it

Ollama
$ ollama run llama4:128x17b
llama.cpp
$ llama-cli -hf unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF:Q4_K_M
LM Studio
$ lms get unsloth/Llama-4-Maverick-17B-128E-Instruct-GGUF

Which devices can run Llama 4 Maverick?

Apple Silicon Macs

RAM-only laptops

iPhone & iPad

Android

NVIDIA GPUs

AMD GPUs

Same job, different size

Pick by what fits your memory: step down to free up VRAM, or step up if you have headroom.

Next size down GLM-4.6 357B Next size up MiniMax M3 428B

Similar models

Head-to-head

FAQ

How much VRAM or RAM does Llama 4 Maverick need?

At Q4_K_M, Llama 4 Maverick needs about 231.7 GB (weights ~226.09 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~402.2 GB.

What is the Q4_K_M GGUF file size of Llama 4 Maverick?

The Q4_K_M GGUF file is about 226.09 GB to download, and the Q8_0 GGUF is about 396.57 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~231.7 GB of memory at Q4_K_M.

Can Llama 4 Maverick run on a laptop?

Llama 4 Maverick is large; you need a high-memory Mac or multi-GPU setup at Q4_K_M.

Is Llama 4 Maverick cheaper to run because it is a MoE model?

It is faster, not lighter. Llama 4 Maverick activates only 17B of 400B params per token (so it runs quickly), but all experts must stay in memory, so it still needs memory for the full 400B.

Can I use Llama 4 Maverick commercially?

Conditionally. Llama 4 Community License; commercial use allowed below Meta's user threshold.

Understand the numbers

Short guides to the ideas behind Llama 4 Maverick's memory and quant figures.

The larger Llama 4: 400B total / 17B active across 128 experts, 1M context. Needs server-class memory even at Q4.

Sources

Last validated 2026-08-03. Memory figures are estimates. See methodology.