{"type": "article", "title": "FP8 MTP draft stack for vLLM — +6.6% decode on modelopt NVFP4 Qwen3.5-family checkpoints", "publisher": "Web Pulse", "url": "https://wpnews.pro/news/fp8-mtp-draft-stack-for-vllm-6-6-decode-on-modelopt-nvfp4-qwen3-5-family", "original_source": "https://gist.github.com/jaderinoo/0d1c0487073cef9d30ef4fd03499fa54", "published": "2026-09-15T17:52:06+00:00", "accessed": "2026-09-30", "id": "fp8-mtp-draft-stack-for-vllm-6-6-decode-on-modelopt-nvfp4-qwen3-5-family"}