Comment by wxw

Thanks for including a section on Token Efficiency (https://x.ai/news/grok-4-5#faster-than-flash-models), hope to see this more prominently in all model releases.