Skip to content
localmodel.run

Text model · llama

Llama 3.2 3B requirements

llama family · 3B params · released Sep 2024 · 59.5M Ollama pulls · LMArena Elo 1166. Minimum to run at Q4_K_M: Apple M1 (8GB).

LicenseLlama 3.2 Community· Conditional↓ 1M/mo♥ 2.2Kon HuggingFace
Q4_K_M
2.02 GB
Q8_0
3.42 GB
Total @ Q4 (4k)
~3.2 GB
Context
128 k

Quantization sizes

GGUF quantson disk
QuantizationSize on disk
Q2_K1.3 GB est
Q3_K_M1.5 GB est
Q4_K_M (default)2.02 GB
Q5_K_M2.1 GB est
Q6_K2.5 GB est
Q8_03.42 GB
FP166 GB est

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

Run it

Ollama
$ ollama run llama3.2:3b
llama.cpp
$ llama-cli -hf unsloth/Llama-3.2-3B-Instruct-GGUF:Q4_K_M
LM Studio
$ lms get unsloth/Llama-3.2-3B-Instruct-GGUF

Which devices can run Llama 3.2 3B?

FAQ

How much VRAM or RAM does Llama 3.2 3B need?

At Q4_K_M, Llama 3.2 3B needs about 3.2 GB (weights ~2.02 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~4.6 GB.

Can Llama 3.2 3B run on a laptop?

Yes, Llama 3.2 3B fits on a 16 GB machine at Q4_K_M and runs on Apple Silicon or a 12 GB+ GPU comfortably.

Can I use Llama 3.2 3B commercially?

Conditionally. Llama 3.2 Community License: free under 700M MAU.

Released September 25, 2024. Text-only edge model. Ollama default tag is 3B at 2.0GB. HF unsloth repo confirms 2.02GB Q4_K_M and 3.42GB Q8_0.

Sources

Memory figures are estimates. See methodology.