I see there's no option for purchasing a Groq LPU or a rack? Is it still possible? Can a single LPU run gpt-oss-120B? How many real-time requests can a single LPU or a RACK handle on average?
Unfortunately, we don’t really sell racks anymore
Hi @yawnxyz , Hope you are doing well. We want to have an on premise solution running. Currently we have system but that is really slow and hard to scale. What are my options, we can’t use cloud service.
Appreciate if you can help and guide us on this.
Thanks
Have a nice day.
Hi @mobi we support dedicated pinning to our data centers through enterprise, but unfortunately we don’t do on-premise racks (we’re bringing up more and more data centers around the world though). Could you describe your use case a bit more?