한국어EN
On-Device & Local Inferencellama.cpp (ggml-org)llama.cppCommunity

llama.cpp merges PR #19769, adding NVFP4

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

Merged PR #19769 — introduced GGML_TYPE_NVFP4. FP4 E2M1 weights + UE4M3 block scales, loading NVIDIA ModelOpt outputs directly

Why it matters

GGUF absorbed a hardware-native FP4 format beyond its own K-quants

Source — Primary sources first; secondary reports are labelled “Press”

Officialgithub.comhttps://github.com/ggml-org/llama.cpp/pull/19769

Added in Week 35, 2026 · See that week

On the map — ⑲ On-Device line

  1. Previous stopUygar Kurt publishes a unified evaluation of llama.cpp quantization
  2. This recordllama.cpp merges PR #19769, adding NVFP4
  3. Next stopGoogle releases Gemma 4 as an AICore developer preview
See it on the map

Same week

Skill Issue: Harness Engineering for Coding AgentsOfficialQdrant announces a $50 million Series BOfficialAMI Labs raises $1.03 billion at a $3.5 billion pre-money valuationPress

Nearby on the same axis

Google releases Gemma 4 as an AICore developer previewOfficialMicrosoft makes Foundry Local generally available (GA)OfficialChrome 148 stable opens the Prompt API to the webOfficialUygar Kurt publishes a unified evaluation of llama.cpp quantizationPaper