nodus get gpus (also nodus get offerings) lists every offering you can run on now: accelerator, count, host
vCPU and memory, region class, the from rate per hour and availability. --gpu, --count, --region and
--interruptible narrow the list, and -o wide adds typical and list rates, startup time and interruption rate.
See Choose GPUs and check availability.
The console’s Runs › GPUs tab shows the same list. Rates are what you pay per machine-hour, never above the list
price, and the rate a run starts at is frozen for that run.