MiniMax open-sources M3.1 Flash: Citi focuses on inference efficiency and continuity of overseas services
M3.1 Flash processed more than 100 trillion tokens in its first 72 hours. Citi attributes its pace of adoption to lower inference cost per unit and improved utility for enterprise agents. Completion of the procedures related to overseas open-sourcing also eased concerns about service disruption, but valuation remains constrained by the risk of customers switching models and risks surrounding new product launches.
Citi ·