{"type":"post","id":"nvidia-has-published-an-nvfp4-quantized-deepseek-v4-a06093","slug":"nvidia-has-published-an-nvfp4-quantized-deepseek-v4-a06093","title":"NVIDIA has published an NVFP4-quantized DeepSeek-V4.1-Flash-1017 checkpoint for Blackwell GPUs","description":"NVIDIA has published an NVFP4-quantized DeepSeek-V4.1-Flash-1017 checkpoint for Blackwell GPUs.","retrieval_nugget":"NVIDIA has published an NVFP4-quantized DeepSeek-V4.1-Flash-1017 checkpoint for Blackwell GPUs. Its model card reports a smaller footprint—about 400 GiB rather than 475 GiB—with scores close to the original on the listed evaluations, but serving it requires a vLLM build with a specific Engram-support change.","published_at":"2026-11-01","updated_at":"2026-11-01","record_date":"2026-11-01","date_kind":"published_at","topics":[],"entities":["NVIDIA","DeepSeek AI","DeepSeek-V4.1-Flash-1017-NVFP4","DeepSeek-V4.1-Flash-1017","vLLM","NVIDIA Model Optimizer"],"related_patterns":[],"novelty_label":"new_fact","what_changed":"NVIDIA published a Blackwell-specific NVFP4 checkpoint, roughly 400 GiB versus 475 GiB with near-identical scores, but it requires patched vLLM.","prior_refs":["memory:9931b90e-7f0f-8aeb-a2b7-daced1c85cb2","memory:8ecc42e7-9646-878f-a3e5-c74ca85a1d06","memory:4c416acd-47a0-818c-af99-bc98dd032139"],"source_url":"https://huggingface.co/nvidia/DeepSeek-V4.1-Flash-1017-NVFP4","source_urls":["https://huggingface.co/nvidia/DeepSeek-V4.1-Flash-1017-NVFP4"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:nvidia-has-published-an-nvfp4-quantized-deepseek-v4-a06093","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/nvidia-has-published-an-nvfp4-quantized-deepseek-v4-a06093/","markdown":"https://newruntime.com/posts/nvidia-has-published-an-nvfp4-quantized-deepseek-v4-a06093.md","json":"https://newruntime.com/posts/nvidia-has-published-an-nvfp4-quantized-deepseek-v4-a06093.json"},"next_reads":[{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/source-ledger.json","url":"https://newruntime.com/source-ledger.json","title":"Source Ledger","reason":"Inspect the public source evidence graph.","media_type":"application/json"}]}