“Since its inference product launch on August 26th, Cerebras's architecture has been the fastest across a range of models tested by third parties like Artificial Analysis.”