MiniMax-M3 is a multimodal foundation model that supports text, image, and video inputs with text output. It features a 1M-token context window and is optimized for long-horizon agentic work, coding, and tool use. Built on MiniMax Sparse Attention (MSA), it achieves significantly faster prefill and decode speeds while maintaining quality across tasks. The model is trained as a native multimodal system and tuned for multi-turn, production-like collaboration.