AI 导读
突发新闻🚨：RUBIN NVL72 在 GLM5.3 744B 上、面向智能体工作负载达到 862 tok/s/user！ 不到 8 天，NVIDIA 通过低延迟专用 kernel 将 Rubin 的交互性提升了近 2 倍。
正文
BREAKING NEWS🚨: RUBIN NVL72 HITS 862 tok/s/user on GLM5.3 744B on agentic workloads!
In less than 8 days, NVIDIA has increased interactivity on Rubin by close to 2x through low-latency specialized kernels.
来源：SemiAnalysis · x.com