Core Insights - MiniMax, a domestic AI unicorn, has launched its new generation text model, MiniMax M2, which is optimized for agent and coding applications [1] Pricing and Performance - The API pricing for the model is set at $0.3 (2.1 RMB) per million tokens for input and $1.2 (8.4 RMB) for output [1] - The model offers an online inference service with a throughput of approximately 100 TPS (tokens per second) [1] Global Ranking and Optimization - According to the latest rankings from global AI benchmarking organization A.jpgicial Analysis, MiniMax M2 ranks among the top globally in core tasks such as coding and agents [1] - MiniMax M2 is specifically optimized for coding and agent tasks, enabling automatic debugging and repair of multiple code files while maintaining lower latency and cost, as well as higher throughput efficiency [1]
MiniMax开源并上线新一代文本大模型
Zhong Zheng Wang·2025-10-27 13:15