NVIDIA Releases NVFP4 Quantized Qwen3.8-Flash-Next on Hugging Face

Aliasghar Khani, an ML researcher at Autodesk who curates AI and ML papers on X under @_akhaliq to roughly 493k followers, retweeted a post from HuggingPapers. That post states NVIDIA just released the NVFP4 quantized Qwen3.8-Flash-Next on Hugging Face. The model is presented as a 125B MoE with hybrid attention now at 63% sma. The retweet forms the visible evidence in the packet and attributes the release claim solely to the shared source rather than confirming it independently.