MiniMax M3

MiniMax M3 is a frontier open-weight foundation model combining native multimodal understanding (text, image, video), strong coding and agentic task execution, and up to 1M-token context via MiniMax Sparse Attention (MSA). It is positioned as the first open-weight model to unify frontier coding performance, 1M context, and native multimodality in a single release. Source: supabase.