DeepSeek kicks off 2026 with new AI architecture aimed at more efficient model training GizmochinaDeepSeek Touts New Training Method as China Pushes AI Efficiency livemint.comDeepSeek opens 2026 with paper signalling race to train bigger models for less South China Morning PostDeepSeek develops mHC AI architecture to boost model performance SiliconANGLEDeepSeek releases new paper by Liang Wenfeng: Proposes mHC new architecture to improve large model training stability Bitget
