Open-weight & owned stack · Quantization efficiency

Nunchux Introduces Low-Bit VC-Attention Kernel

Sep 17, 2026, first seen via MarkTechPost

Nunchux introduced a training-free low-bit attention kernel that it says accelerates video diffusion and MiniMax-H3 workloads on B200 GPUs.

Read the original at marktechpost.comOpens the publisher's site in a new tab

Also covering this

Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers - MarkTechPostmarktechpost.com, Sep 17

More in Open-weight & owned stack

Xiaomi Publishes MiMo-V2.6 Reinforcement Learning DashboardSep 17 Gemma Models Surpass One Billion DownloadsSep 16 llama.cpp Fixes Remote-Code-Execution VulnerabilitySep 16 Alibaba Releases Occamy Open-Weight AI AgentSep 15 Apple Enables Local Foundation Models On MacOSSep 15