LMSYS：Blog（Chatbot Arena 团队）·· 13 小时前AI 评分44
SGLang 在 NVIDIA GB300 NVL72 上实现最高 25 倍推理性能提升
Blog Unlocking 25x Inference Performance with SGLang on NVIDIA GB300 NVL72 The SGLang team has worked closely with NVIDIA across multiple GPU generations to unlock step-function gains in inference performance for large-scale deployments of Mixture of Expert (MoE) reasoning m... NVIDIA and Community SGLang Developers February 20, 2026
AI 导读
SGLang 与 NVIDIA 合作在 GB300 NVL72 上实现最高 25 倍推理性能提升，基于 SemiAnalysis InferenceXv2 运行 DeepSeek R1 对比 H200（50 TPS/user 延迟约束）。
来源：LMSYS：Blog（Chatbot Arena 团队） · lmsys.org