7 ms·MiniMax-M1, first open-weight, large-scale hybrid-attention reasoning model3 points by hallh 1y ago