@TeksEdge
π’Huge Open-Source Drop! MiniMax just released M3 Weights π₯ This is the one many of us are waiting for! β 428B total parameters MoE β Only ~23B active parameters (extremely sparse) β Advanced Sparse Attention β up to 1M context β Strong in coding & agentic tasks β Fully open-weights on Hugging Face π― This is one of the most efficient frontier-level models released so far. π₯ Big Win for local AI & open source community π