minimax-m3.1-flash

MiniMax M3.1 Flash is a new-generation high-speed multimodal foundation model developed by MiniMax. It features a 1M-token ultra-long context window and tunable thinking depth, focusing on agentic reasoning, coding engineering and complex task execution. It supports natural language conversation, image recognition, deep logical reasoning, long-document processing and text-image multimodal fusion. The model delivers greatly improved inference speed with reduced API cost, suitable for software dev