--- format: "aidr-story-markdown/v1" id: "f67cf667c292b6866817f8675b21c37cf4c6ac1b0ee4ed375074e1ef0173d728" canonical_url: "https://aidr.today/f67cf667?lang=en" title: "SGLang Launches New Plugin System for NVIDIA and AMD AI Kernels" lang: "en" requested_lang: "en" available_langs: ["en","vi"] translation_fallback: null fallback_fields: [] published_at: "2026-10-20T17:40:15.000Z" category: "Frameworks" topics: ["deepseek","inference","framework"] source_urls: ["https://huggingnews.com/ai/update-sglang-launches-new-plugin-system-for-nvidia-and-amd-ai-kernels-f9948f1c","https://marketbrief.now/ai/update-sglang-launches-new-plugin-system-for-nvidia-and-amd-ai-kernels-f9948f1c"] summary: "SGLang has launched a new plugin system using two extensions, sgl-nvidia-mega and sgl-amd-mono, to allow hardware vendors to deliver low-latency kernels as external libraries. The NVIDIA plugin is currently limited to DeepSeek-V4.1-Flash running on 4x GB300 GPUs. The system is already enabling third-party optimizations, such as a Zyphra kernel that increased single-user decode speeds at 1M context from 43.5 to 61.1 tok/s for DeepSeek-V4. SGLang is developing additional kernels for GLM-5.3, K3, and MiniMax M3 models for both NVIDIA and AMD hardware." --- # SGLang Launches New Plugin System for NVIDIA and AMD AI Kernels > [Open the canonical story]() **Published:** 2026-10-20T17:40:15.000Z **Category:** Frameworks **Topics:** deepseek, inference, framework ## Summary SGLang has launched a new plugin system using two extensions, sgl\-nvidia\-mega and sgl\-amd\-mono, to allow hardware vendors to deliver low\-latency kernels as external libraries\. The NVIDIA plugin is currently limited to DeepSeek\-V4\.1\-Flash running on 4x GB300 GPUs\. The system is already enabling third\-party optimizations, such as a Zyphra kernel that increased single\-user decode speeds at 1M context from 43\.5 to 61\.1 tok/s for DeepSeek\-V4\. SGLang is developing additional kernels for GLM\-5\.3, K3, and MiniMax M3 models for both NVIDIA and AMD hardware\. ## Sources - [Story source]() - [Story source]()