AI 导读
NVIDIA 在其 GLM 5.3 SGLang 对比中，在 150 tok/s/user 下成本效率最高可达 AMD 的 5 倍。 - 来自最新 SemiAnalysis 报告 “以 AMD 软件当前的状态，在 150 tok/s/user 下，Nvidia 的性能优势如此之大，以至于即使竞争对手的芯片硬件免费赠送（但服务商仍需支付数据中心托管、电力及其他运营成本），使用 Nvidia 时每 token 成本仍然更低。”
正文
NVIDIA reaches up to 5x better cost efficiency than AMD at 150 tok/s/user on its GLM 5.3 SGLang comparison.
- Latest SemiAnalysis report
"With the current state of AMD software, at 150 tok/s/user, Nvidia’s performance advantage is so great that even if the competitor chip hardware was sold for free (but with providers still of course paying for datacenter hosting and power and other operating costs), cost per token would still be cheaper when using Nvidia."
来源：@rohanpaul_ai · x.com