MiniMaxMiniMax M3.1 Flash API

Fast, low-cost MiniMax model — always-on reasoning with a 1M-token context.