Llama 3.2 3B

small

Fast and lightweight. Good for quick tasks and low-resource devices.

Parameters3.0B
Context131,072 tokens
Download2.0 GB
Min RAM4 GB
Architecturellama

Available quantizations

QuantSizeQuality
Q4_K_M2.0 GBgood
Q5_K_M2.3 GBbetter
Q8_03.4 GBbest

Check whether your machine can run this model — with real measured speeds from anonymous community benchmarks — on the Central-Intel compatibility page.