Mistral AI

Ministral 3

small machinesvision

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.

3B

8 GB card
Parameters
3,8B
Q4_K_M file
2,0 Go
Recommended VRAM
≥ 3,5 Go
Max context
262 144

Not measured yet

8B

8 GB card
Parameters
8,9B
Q4_K_M file
4,8 Go
Recommended VRAM
≥ 6,3 Go
Max context
262 144

Not measured yet

14B

12 GB card
Parameters
13,9B
Q4_K_M file
7,7 Go
Recommended VRAM
≥ 9,2 Go
Max context
262 144

Not measured yet

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).