Home / Will it run? / Llama 3.2 3B Instruct
Meta · Llama 3.2
Will Llama 3.2 3B Instruct run on my GPU?
3.21B parameters, 131,072-token context, Llama 3.2 Community License. At Q4_K_M the weights come to about 1.8 GiB; add 1.5 GiB for the runtime and you need roughly 3.3 GiB of VRAM before context.
Weights by quant
GiBQ4_K_M · the everyday quant1.8 GiB
Q8_0 · near-lossless3.2 GiB
FP16 · full weights6.0 GiB
Runtime reserve1.5 GiB
Card by card
18 of 18 fit at Q4Headroom is what remains for the KV cache after weights and the 1.5 GiB reserve. How far it stretches depends on the architecture: grouped-query models are frugal, older dense models are not. If the table says tight, plan on a shorter context or a lower quant. Parameter count from the model's own card. The Buy column opens live Amazon listings; those links may earn us a commission.