croes on Cerebras Inference now 3x faster: Llama3.1-70B breaks 2,100 tokens/s · Hackerly

Comment on Cerebras Inference now 3x faster: Llama3.1-70B breaks 2,100 tokens/s parent

It has a waiting list