← Glossary Model
H200
A high-end Nvidia data centre GPU used for training and running AI models, the memory-upgraded successor to the H100.
The H200 is one of Nvidia’s data centre graphics processors, the kind that lives in a server rack rather than a desktop. It is the successor to the widely used H100, with the main upgrade being considerably more and faster memory, which matters because a large model has to fit in memory to run at all.
You will see it quoted as a unit of measurement. When a lab says a model transcribes three hours of audio in a second “on a single H200”, that is shorthand for a specific and expensive piece of hardware, and it is worth remembering before assuming the same speed on your own machine.
Mentioned in