Comment by FezzikTheGiant

what's the cost difference between groq/cerebras vs using something else for inferencing open source models? I'm guessing the speed comes at a cost?