NVIDIA launched an AI ChatBot which can run locally

[harath]

But inorder to use it we need to have 30 or 40 series of their GPU ...

[prnva_]

Its 35gb, requires 8gb vram and ofcourse rtx 30/40 series gpu.

Another alternative is llamafile by Mozilla but its quite slower.

[DK-oxze]

And how much RAM and GPU compute it needs to run?