MiniMax teases upcoming M3 model with new sparse attention mechanism and 15.6X long-context response speed boost

It directly solves the exact bottleneck that normally makes AI chatbots freeze or stutter when handling massive amounts of ...