Meta

Llama 3.2

small machines

Pick a size: the video memory shown is the Q4_K_M file (the protocol's quantisation) plus about 1.5 GB for a 4,000-token context.

1B

8 GB card
Parameters
1,2B
Q4_K_M file
0,8 Go
Recommended VRAM
≥ 2,3 Go
Max context
131 072

Not measured yet

3B

8 GB card
Parameters
3,2B
Q4_K_M file
1,9 Go
Recommended VRAM
≥ 3,4 Go
Max context
131 072

Measured on 2 cards · 270 tok/s (RTX 5070 Ti)

MoE (“30B-A3B”): all the memory of a 30B, but the speed of a model with 3B active parameters. “> 32 GB”: several cards, or part of the model in system memory (much slower).