Cerebras reports 981 tokens per second on Kimi K2.6 model, 6.7x faster than GPU cloud

Comments

Join the discussion on this story.