We provide a mix of latest‑gen GPUs tuned for training and inference. Below are sample instance types — final availability depends on region and capacity.
